openapi: 3.1.0 info: title: Gateway REST API version: 5.10.0 tags: - name: gateway.openapi_AccountService x-displayName: AccountService - name: gateway.openapi_DeploymentService x-displayName: DeploymentService - name: gateway.openapi_Gateway x-displayName: Gateway - name: gateway.openapi_ModelService x-displayName: ModelService - name: gateway.openapi_TrainingService x-displayName: TrainingService - name: gateway-extra.openapi_Gateway x-displayName: Gateway - name: responses-harmonized.openapi_other x-displayName: other - name: text-completion.openapi_other x-displayName: other - name: anthropic-messages.openapi_other x-displayName: other paths: /v1/accounts: servers: - url: https://api.fireworks.ai get: summary: List Accounts operationId: Gateway_ListAccounts responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListAccountsResponse" parameters: - name: pageSize description: >- The maximum number of accounts to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListAccounts call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListAccounts must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only accounts satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: |- Not supported. Accounts will be returned ordered by `name`. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/auth/refresh: servers: - url: https://api.fireworks.ai post: summary: Refresh a session JWT before expiry with a fresh TTL. operationId: Gateway_RefreshSessionToken responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRefreshSessionTokenResponse" requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayRefreshSessionTokenRequest" description: Request to refresh an existing session JWT before it expires. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/creditCodes:redeem: servers: - url: https://api.fireworks.ai post: summary: |- APIs for credit codes. Redeem Credit Code operationId: Gateway_RedeemCreditCode responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRedeemCreditCodeResponse" requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayRedeemCreditCodeRequest" description: >- RedeemCreditCodeRequest is the request to redeem a credit code for an account. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/validateModelConfig: servers: - url: https://api.fireworks.ai post: summary: Validate Model Config operationId: Gateway_ValidateModelConfig responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayValidateModelConfigRequest" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}: servers: - url: https://api.fireworks.ai get: summary: Get Account operationId: Gateway_GetAccount responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayAccount" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/accountUsageFilterOptions: servers: - url: https://api.fireworks.ai get: summary: >- Distinct filter values for an account/time range (serverless + dedicated; for FE, separate -yet mirrors GetAccountUsage). operationId: Gateway_GetAccountUsageFilterOptions responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetAccountUsageFilterOptionsResponse parameters: - name: startTime in: query required: true schema: type: string format: date-time - name: endTime in: query required: true schema: type: string format: date-time - name: usageType description: >- If not specified, loads filter options for serverless and dedicated usage only. - SERVERLESS: Serverless filter dimensions only (model_name, api_key_id, api_key_name, user_id, session_id, annotations.*). - DEDICATED_DEPLOYMENT: Dedicated deployment filter dimensions (deployment_name, annotations.team, .project, .environment). - TRAINING: Training filter dimensions (job_id, training_session_id, job_type, usage_type, accelerator_type, base_model). in: query required: false schema: type: string enum: - USAGE_TYPE_UNSPECIFIED - SERVERLESS - DEDICATED_DEPLOYMENT - TRAINING default: USAGE_TYPE_UNSPECIFIED - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/auditLogs: servers: - url: https://api.fireworks.ai get: summary: List User Audit Logs operationId: Gateway_ListAuditLogs responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListAuditLogsResponse" parameters: - name: startTime description: |- Start time of the audit logs to retrieve. If unspecified, the default is 30 days before now. in: query required: false schema: type: string format: date-time - name: endTime description: |- End time of the audit logs to retrieve. If unspecified, the default is the current time. in: query required: false schema: type: string format: date-time - name: email description: |- Optional. Filter audit logs for user email associated with the account. in: query required: false schema: type: string - name: pageSize description: >- The maximum number of audit logs to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 10. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListAuditLogs call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListAuditLogs must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: Unused but required to use existing ListRequest functionality. in: query required: false schema: type: string - name: orderBy description: Unused but required to use existing ListRequest functionality. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/batchInferenceJobs: servers: - url: https://api.fireworks.ai get: summary: List Batch Inference Jobs operationId: Gateway_ListBatchInferenceJobs responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListBatchInferenceJobsResponse" parameters: - name: pageSize description: >- The maximum number of batch inference jobs to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListBatchInferenceJobs call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListBatchInferenceJobs must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only jobs satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "created_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Batch Inference Job operationId: Gateway_CreateBatchInferenceJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayBatchInferenceJob" parameters: - name: batchInferenceJobId description: ID of the batch inference job. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayBatchInferenceJob" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/batchInferenceJobs/{batch_inference_job_id}: servers: - url: https://api.fireworks.ai get: summary: Get Batch Inference Job operationId: Gateway_GetBatchInferenceJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayBatchInferenceJob" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: batch_inference_job_id in: path required: true description: The Batch Inference Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Batch Inference Job operationId: Gateway_DeleteBatchInferenceJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: batch_inference_job_id in: path required: true description: The Batch Inference Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/billing/summary: servers: - url: https://api.fireworks.ai get: summary: Get billing summary information for an account operationId: Gateway_GetBillingSummary responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetBillingSummaryResponse" parameters: - name: startTime description: >- Start time for the billing period. Note: Costs are aggregated daily. Only the date portion (YYYY-MM-DD) is used; the time portion is ignored. For example, 2025-10-05T07:18:29Z and 2025-10-05T23:59:59Z are treated the same as 2025-10-05T00:00:00Z. in: query required: true schema: type: string format: date-time - name: endTime description: >- End time for the billing period (exclusive). Note: Costs are aggregated daily. Only the date portion (YYYY-MM-DD) is used; the time portion is ignored. Costs for the end date are NOT included. For example, to get costs for Oct 5 and Oct 6, use: start_time: 2025-10-05T00:00:00Z end_time: 2025-10-07T00:00:00Z (Oct 7 is excluded) in: query required: true schema: type: string format: date-time - name: granularity description: >- Requested breakdown granularity. Defaults to UNSPECIFIED (no breakdown). in: query required: false schema: type: string enum: - GRANULARITY_UNSPECIFIED - DAILY default: GRANULARITY_UNSPECIFIED - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/usageCosts:query: servers: - url: https://api.fireworks.ai post: summary: Query grouped usage cost subtotals for an account. description: >- Returns rated dollar subtotals for usage over a time range, grouped by up to two of: HOUR, DAY, MODEL, USER, or API_KEY. Unlike `GET /billingUsage` (metered quantities) and `GET /billing/summary` (line items by billing category), this endpoint returns *rated costs* broken down by caller-supplied dimensions, with pagination and an account-wide `subtotal`. Requires account administrator access for `ACCOUNT` scope; `SELF` scope returns costs for the authenticated principal only. operationId: Gateway_QueryUsageCosts responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayQueryUsageCostsResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayQueryUsageCostsBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/billingSettings: servers: - url: https://api.fireworks.ai get: summary: Returns the singleton BillingSettings for the given account. operationId: Gateway_GetBillingSettings responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayBillingSettings" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Updates the singleton BillingSettings for the given account. operationId: Gateway_UpdateBillingSettings responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayBillingSettings" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: type: object properties: updateTime: type: string format: date-time description: The time the billing settings were last updated. readOnly: true enableCreditAutoReload: type: boolean description: If true, the account will auto-reload its credit balance. creditAutoReloadThreshold: $ref: "#/components/schemas/typeMoney" description: >- The threshold at which to auto-reload the account's credit balance. Must be less than $100,000. creditAutoReloadAmount: $ref: "#/components/schemas/typeMoney" description: |- The amount to auto-reload the account's credit balance. Must be less than $100,000. title: >- The billing settings to update. `billing_settings.name` must be populated. Format: accounts/{account}/billingSettings description: >- The billing settings to update. `billing_settings.name` must be populated. Format: accounts/{account}/billingSettings required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/billingUsage: servers: - url: https://api.fireworks.ai get: summary: Get Account Usage operationId: Gateway_GetAccountUsage responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayAccountUsage" parameters: - name: startTime description: |- Costs returned are inclusive of `start_time`. start_time must be before end_time. in: query required: true schema: type: string format: date-time - name: endTime description: |- Costs returned are exclusive of `end_time`. end_time must not be more than 31 days after start_time. in: query required: true schema: type: string format: date-time - name: usageType description: >- Usage type to query usage for If not specified, returns all usage types (serverless, dedicated deployments, and training). - USAGE_TYPE_UNSPECIFIED: Default value. When specified (or when usage_type field is not set), returns usage data for all deployment types: serverless requests, dedicated deployments, and training jobs. - SERVERLESS: Returns only serverless usage data. Filters the response to include only usage from serverless API requests. - DEDICATED_DEPLOYMENT: Returns only dedicated deployment usage data. Filters the response to include only usage from dedicated deployments. - TRAINING: Returns only training job usage data (SFT/DPO token usage and RFT / service-mode trainer GPU-seconds usage). Inference deployments serving rollouts for RFT / online RL are reported under DEDICATED_DEPLOYMENT (not TRAINING) to avoid double counting GPU time. in: query required: false schema: type: string enum: - USAGE_TYPE_UNSPECIFIED - SERVERLESS - DEDICATED_DEPLOYMENT - TRAINING default: USAGE_TYPE_UNSPECIFIED - name: timezone description: >- IANA timezone identifier for daily aggregation (e.g., "America/Los_Angeles", "Europe/London"). When specified, the returned data will be aggregated into daily buckets based on this timezone. If not specified or empty, defaults to "UTC". See: https://en.wikipedia.org/wiki/List_of_tz_database_time_zones in: query required: false schema: type: string - name: groupBy description: >- Dimensions to group usage by (multiple values allowed; each is a separate GROUP BY column). Serverless: "model_name", "api_key_id", "api_key_name", "user_id", "session_id", "annotations.team", "annotations.project", "annotations.environment". Dedicated: "deployment_name", "accelerator_type", and the same annotation keys. Training: "job_id", "training_session_id", "job_type", "usage_type", "accelerator_type", "base_model", and the same annotation keys. When usage_type is unspecified, dimensions that apply only to one stream are ignored on the others (e.g. "deployment_name" is ignored for serverless and training; "model_name" / "api_key_id" / "api_key_name" / "user_id" / "session_id" are ignored for dedicated and training; "job_id" / "training_session_id" / "job_type" are ignored for serverless and dedicated). "session_id" is the opaque client session from billing properties (serverless-only, high cardinality): prefer filtering by a single session; grouping by it can fan out to one row per session. Example: ["annotations.team", "model_name"] or ["user_id", "model_name"] or ["api_key_id", "api_key_name"]. If empty: serverless aggregates by model name; dedicated defaults to deployment and accelerator type; training aggregates by job_id, training_session_id, job_type, usage_type, accelerator_type and base_model. in: query required: false explode: true schema: type: array items: type: string - name: filter description: >- This is a request variable of the map type. The query format is "map_name[key]=value", e.g. If the map name is Age, the key type is string, and the value type is integer, the query parameter is expressed as Age["bob"]=18 in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/checkpoints/{checkpoint_id}:promote: servers: - url: https://api.fireworks.ai post: summary: >- Promote a checkpoint to a model. The checkpoint is identified by account + snapshot ID; the trainer job ID is passed in the request body to resolve the GCS bucket. operationId: Gateway_PromoteCheckpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPromoteCheckpointResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: checkpoint_id in: path required: true description: The Checkpoint Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayPromoteCheckpointBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/clusters: servers: - url: https://api.fireworks.ai get: summary: List Clusters operationId: Gateway_ListClusters responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListClustersResponse" parameters: - name: pageSize description: >- The maximum number of clusters to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListClusters call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListClusters must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only clusters satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Cluster operationId: Gateway_CreateCluster responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayCluster" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateClusterBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/clusters/{cluster_id}: servers: - url: https://api.fireworks.ai get: summary: Get Cluster operationId: Gateway_GetCluster responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayCluster" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: cluster_id in: path required: true description: The Cluster Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Cluster operationId: Gateway_UpdateCluster responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayCluster" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: cluster_id in: path required: true description: The Cluster Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: >- Human-readable display name of the cluster. e.g. "My Cluster" Must be fewer than 64 characters long. createTime: type: string format: date-time description: The creation time of the cluster. readOnly: true eksCluster: $ref: "#/components/schemas/gatewayEksCluster" fakeCluster: $ref: "#/components/schemas/gatewayFakeCluster" state: $ref: "#/components/schemas/gatewayClusterState" description: The current state of the cluster. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: >- Detailed information about the current status of the cluster. readOnly: true updateTime: type: string format: date-time description: The update time for the cluster. readOnly: true title: |- The properties of the cluster being updated. `cluster.name` must be populated with the updated resource's name. description: |- The properties of the cluster being updated. `cluster.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Cluster operationId: Gateway_DeleteCluster responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: cluster_id in: path required: true description: The Cluster Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/clusters/{cluster_id}:getConnectionInfo: servers: - url: https://api.fireworks.ai get: summary: Get Cluster Connection Info description: Retrieve connection settings for the cluster to be put in kubeconfig operationId: Gateway_GetClusterConnectionInfo responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayClusterConnectionInfo" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: cluster_id in: path required: true description: The Cluster Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/creditRedemptions: servers: - url: https://api.fireworks.ai get: summary: List Credit Redemptions operationId: Gateway_ListCreditRedemptions responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListCreditRedemptionsResponse" parameters: - name: pageSize description: The maximum number of redemptions to return in: query required: false schema: type: integer format: int32 - name: pageToken description: A page token, received from a previous ListCreditRedemptions call in: query required: false schema: type: string - name: filter description: Filter string to filter redemptions in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets: servers: - url: https://api.fireworks.ai get: summary: List Datasets operationId: Gateway_ListDatasets responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDatasetsResponse" parameters: - name: pageSize description: >- The maximum number of datasets to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDatasets call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDatasets must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only model satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Dataset operationId: Gateway_CreateDataset responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDataset" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateDatasetBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets/{dataset_id}: servers: - url: https://api.fireworks.ai get: summary: Get Dataset operationId: Gateway_GetDataset responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDataset" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Dataset operationId: Gateway_UpdateDataset responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDataset" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string createTime: type: string format: date-time readOnly: true state: $ref: "#/components/schemas/gatewayDatasetState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true exampleCount: type: string format: int64 userUploaded: $ref: "#/components/schemas/gatewayUserUploaded" evaluationResult: $ref: "#/components/schemas/gatewayEvaluationResult" transformed: $ref: "#/components/schemas/gatewayTransformed" splitted: $ref: "#/components/schemas/gatewaySplitted" evalProtocol: $ref: "#/components/schemas/gatewayEvalProtocol" externalUrl: type: string title: The external URI of the dataset. e.g. gs://foo/bar/baz.jsonl format: $ref: "#/components/schemas/DatasetFormat" createdBy: type: string description: >- The email address of the user who initiated this fine-tuning job. readOnly: true updateTime: type: string format: date-time description: The update time for the dataset. readOnly: true sourceJobName: type: string description: >- The resource name of the job that created this dataset (e.g., batch inference job). Used for lineage tracking to understand dataset provenance. estimatedTokenCount: type: string format: int64 description: The estimated number of tokens in the dataset. readOnly: true averageTurnCount: type: number format: float description: >- An estimate of the average number of turns per sample in the dataset. readOnly: true encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: >- CMEK encryption state (authoritative, immutable, stamped at creation). readOnly: true title: |- The properties of the Dataset being updated. `dataset.name` must be populated with the updated resource's name. description: |- The properties of the Dataset being updated. `dataset.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Dataset operationId: Gateway_DeleteDataset responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets/{dataset_id}:getDownloadEndpoint: servers: - url: https://api.fireworks.ai get: summary: Get Dataset Download Endpoint operationId: Gateway_GetDatasetDownloadEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetDatasetDownloadEndpointResponse" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: downloadLineage description: |- If true, downloads entire lineage chain (all related datasets). Filenames will be prefixed with dataset IDs to avoid collisions. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets/{dataset_id}:getUploadEndpoint: servers: - url: https://api.fireworks.ai post: summary: Get Dataset Upload Endpoint operationId: Gateway_GetDatasetUploadEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetDatasetUploadEndpointResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayGetDatasetUploadEndpointBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets/{dataset_id}:splitDataset: servers: - url: https://api.fireworks.ai post: operationId: Gateway_SplitDataset responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySplitDatasetResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewaySplitDatasetBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets/{dataset_id}:validateUpload: servers: - url: https://api.fireworks.ai post: summary: Validate Dataset Upload operationId: Gateway_ValidateDatasetUpload responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dataset_id in: path required: true description: The Dataset Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayValidateDatasetUploadBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployedModels: servers: - url: https://api.fireworks.ai get: summary: List LoRAs operationId: Gateway_ListDeployedModels responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDeployedModelsResponse" parameters: - name: pageSize description: >- The maximum number of deployed models to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDeployedModels call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDeployedModels must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Only depoyed models satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Load LoRA operationId: Gateway_CreateDeployedModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployedModel" parameters: - name: replaceMergedAddon description: >- Merges new addon to the base model, while unmerging/deleting any existing addon in the deployment. Must be specified for hot reload deployments in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayDeployedModel" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployedModels/{deployed_model_id}: servers: - url: https://api.fireworks.ai get: summary: Get LoRA operationId: Gateway_GetDeployedModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployedModel" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployed_model_id in: path required: true description: The Deployed Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update LoRA operationId: Gateway_UpdateDeployedModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployedModel" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployed_model_id in: path required: true description: The Deployed Model Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: type: string description: Description of the resource. createTime: type: string format: date-time description: The creation time of the resource. readOnly: true model: type: string title: |- The resource name of the model to be deployed. e.g. accounts/my-account/models/my-model deployment: type: string description: >- The resource name of the base deployment the model is deployed to. default: type: boolean description: >- If true, this is the default target when querying this model without the `#` suffix. The first deployment a model is deployed to will have this field set to true. state: $ref: "#/components/schemas/gatewayDeployedModelState" description: The state of the deployed model. readOnly: true serverless: type: boolean title: True if the underlying deployment is managed by Fireworks status: $ref: "#/components/schemas/gatewayStatus" description: Contains model deploy/undeploy details. readOnly: true public: type: boolean description: If true, the deployed model will be publicly reachable. updateTime: type: string format: date-time description: The update time for the deployed model. readOnly: true title: >- The properties of the deployed model being updated. `deployed_model.name` must be populated with the updated resource's name. description: >- The properties of the deployed model being updated. `deployed_model.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Unload LoRA operationId: Gateway_DeleteDeployedModel responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployed_model_id in: path required: true description: The Deployed Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deploymentShapes: servers: - url: https://api.fireworks.ai get: summary: List Deployment Shapes operationId: Gateway_ListDeploymentShapes responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDeploymentShapesResponse" parameters: - name: pageSize description: >- The maximum number of deployments to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDeploymentShapes call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDeploymentShapes must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Only deployment satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "create_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: targetModel description: >- Target model that the returned deployment shapes should be compatible with. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: |- CRUD APIs for deployment shape. Create Deployment Shape operationId: Gateway_CreateDeploymentShape responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeploymentShape" parameters: - name: deploymentShapeId description: >- The ID of the deployment shape. If not specified, a random ID will be generated. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayDeploymentShape" description: The properties of the deployment shape being created. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deploymentShapes/{deployment_shape_id}: servers: - url: https://api.fireworks.ai get: summary: Get Deployment Shape operationId: Gateway_GetDeploymentShape responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeploymentShape" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: skipShapeValidation description: |- If true, returns the latest version regardless of validation status. By default, returns the latest validated version. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_shape_id in: path required: true description: The Deployment Shape Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Deployment Shape operationId: Gateway_UpdateDeploymentShape responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeploymentShape" parameters: - name: fromLatestValidated description: >- When true, the update will use the latest validated version snapshot as the base for fields not present in the update mask; otherwise, the current shape is used. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_shape_id in: path required: true description: The Deployment Shape Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: >- Human-readable display name of the deployment shape. e.g. "My Deployment Shape" Must be fewer than 64 characters long. description: type: string description: >- The description of the deployment shape. Must be fewer than 1000 characters long. createTime: type: string format: date-time description: The creation time of the deployment shape. readOnly: true updateTime: type: string format: date-time description: The update time for the deployment shape. readOnly: true baseModel: type: string description: >- Mutable, but only via UpdateDeploymentShape and only to a compatible model: one with the same model_type, parameter_count, and embedding status. The derived model_type / parameter_count fields (OUTPUT_ONLY, IMMUTABLE) are therefore preserved across a swap. Any incompatible change is rejected. title: >- The base model name. e.g. accounts/fireworks/models/falcon-7b modelType: type: string description: The model type of the base model. readOnly: true parameterCount: type: string format: int64 description: The parameter count of the base model . readOnly: true acceleratorCount: type: integer format: int32 description: >- The number of accelerators used per replica. If not specified, the default is the estimated minimum required by the base model. acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: |- The type of accelerator to use. If not specified, the default is NVIDIA_A100_80GB. precision: $ref: "#/components/schemas/DeploymentPrecision" description: The precision with which the model should be served. disableDeploymentSizeValidation: type: boolean description: If true, the deployment size validation is disabled. enableAddons: type: boolean description: >- If true, LORA addons are enabled for deployments created from this shape. Deprecated: set enable_addons on the deployment instead. draftTokenCount: type: integer format: int32 description: >- The number of candidate tokens to generate per step for speculative decoding. Default is the base model's draft_token_count. draftModel: type: string description: >- The draft model name for speculative decoding. e.g. accounts/fireworks/models/my-draft-model If empty, speculative decoding using a draft model is disabled. Default is the base model's default_draft_model. Deprecated: set default_draft_model on the base model instead. ngramSpeculationLength: type: integer format: int32 description: >- The length of previous input sequence to be considered for N-gram speculation. disableSpeculativeDecoding: type: boolean description: >- DEPRECATED: This field is a no-op. Speculative decoding is configured on deployment. enableSessionAffinity: type: boolean description: Whether to apply sticky routing based on `user` field. numLoraDeviceCached: type: integer format: int32 title: How many LORA adapters to keep on GPU side for caching maxContextLength: type: integer format: int32 description: >- The maximum context length supported by the model (context window). If set to 0 or not specified, the model's default maximum context length will be used. presetType: $ref: "#/components/schemas/DeploymentShapePresetType" description: >- Type of deployment shape for different deployment configurations. title: >- The properties of the deployment shape being updated. `deployment_shape.name` must be populated with the updated resource's name. required: - baseModel description: >- The properties of the deployment shape being updated. `deployment_shape.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Deployment Shape operationId: Gateway_DeleteDeploymentShape responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_shape_id in: path required: true description: The Deployment Shape Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deploymentShapes/{deployment_shape_id}/versions: servers: - url: https://api.fireworks.ai get: summary: List Deployment Shapes Versions operationId: Gateway_ListDeploymentShapeVersions responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListDeploymentShapeVersionsResponse parameters: - name: pageSize description: >- The maximum number of deployment shape versions to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDeploymentShapeVersions call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDeploymentShapeVersions must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Only deployment shape versions satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "create_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_shape_id in: path required: true description: The Deployment Shape Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deploymentShapes/{deployment_shape_id}/versions/{version_id}: servers: - url: https://api.fireworks.ai get: summary: Get Deployment Shape Version operationId: Gateway_GetDeploymentShapeVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeploymentShapeVersion" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_shape_id in: path required: true description: The Deployment Shape Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Deployment Shape Version operationId: Gateway_UpdateDeploymentShapeVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeploymentShapeVersion" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_shape_id in: path required: true description: The Deployment Shape Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string requestBody: content: application/json: schema: type: object properties: createTime: type: string format: date-time description: >- The creation time of the deployment shape version. Lists will be ordered by this field. readOnly: true snapshot: $ref: "#/components/schemas/gatewayDeploymentShape" description: Full snapshot of the Deployment Shape at this version. readOnly: true validated: type: boolean description: If true, this version has been validated. public: type: boolean description: If true, this version will be publicly readable. latestValidated: type: boolean description: >- If true, this version is the latest validated version. Only one version of the shape can be the latest validated version. readOnly: true capabilities: type: array items: $ref: "#/components/schemas/DeploymentShapeVersionCapability" description: The capabilities supported by this deployment shape version. title: >- The properties of the deployment shape version being updated. `deployment_shape_version.name` must be populated with the updated resource's name. description: >- The properties of the deployment shape version being updated. `deployment_shape_version.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments: servers: - url: https://api.fireworks.ai get: summary: List Deployments operationId: Gateway_ListDeployments responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDeploymentsResponse" parameters: - name: pageSize description: >- The maximum number of deployments to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDeployments call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDeployments must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Only deployment satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "create_time". in: query required: false schema: type: string - name: showDeleted description: If set, DELETED deployments will be included. in: query required: false schema: type: boolean - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Deployment operationId: Gateway_CreateDeployment responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployment" parameters: - name: disableAutoDeploy description: >- By default, a deployment created with a currently undeployed base model will be deployed to this deployment. If true, this auto-deploy function is disabled. in: query required: false schema: type: boolean - name: disableSpeculativeDecoding description: >- By default, a deployment will use the speculative decoding settings from the base model. If true, this will disable speculative decoding. in: query required: false schema: type: boolean - name: deploymentId description: >- The ID of the deployment. If not specified, a random ID will be generated. in: query required: false schema: type: string - name: validateOnly description: >- If true, this will not create the deployment, but will return the deployment that would be created. in: query required: false schema: type: boolean - name: skipShapeValidation description: >- By default, a deployment will ensure the deployment shape provided is validated. If true, we will not require the deployment shape to be validated. in: query required: false schema: type: boolean - name: skipImageTagValidation description: >- If true, skip the image tag policy validation that blocks certain image tags. This allows creating deployments with image tags that would otherwise be blocked. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayDeployment" description: The properties of the deployment being created. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}: servers: - url: https://api.fireworks.ai get: summary: Get Deployment operationId: Gateway_GetDeployment responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployment" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Deployment operationId: Gateway_UpdateDeployment responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployment" parameters: - name: skipShapeValidation description: >- By default, updating a deployment shape will ensure the new deployment shape provided is validated. If true, we will not require the deployment shape to be validated. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: >- Human-readable display name of the deployment. e.g. "My Deployment" Must be fewer than 64 characters long. description: type: string description: Description of the deployment. createTime: type: string format: date-time description: The creation time of the deployment. readOnly: true expireTime: type: string format: date-time description: >- Deprecated: This field is deprecated and no longer causes auto-deletion. The time at which this deployment will automatically be deleted. purgeTime: type: string format: date-time description: The time at which the resource will be hard deleted. readOnly: true deleteTime: type: string format: date-time description: The time at which the resource will be soft deleted. readOnly: true state: $ref: "#/components/schemas/gatewayDeploymentState" description: The state of the deployment. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: >- Detailed status information regarding the most recent operation. readOnly: true annotations: type: object additionalProperties: type: string description: >- Annotations to identify deployment properties. Key/value pairs may be used by external tools or other services. The "image-tag-reason" key is redacted from API responses for non-superuser principals. minReplicaCount: type: integer format: int32 description: |- The minimum number of replicas. If not specified, the default is 0. maxReplicaCount: type: integer format: int32 description: |- The maximum number of replicas. If not specified, the default is max(min_replica_count, 1). May be set to 0 to downscale the deployment to 0. maxWithRevocableReplicaCount: type: integer format: int32 description: >- max_with_revocable_replica_count is max replica count including revocable capacity. The max revocable capacity will be max_with_revocable_replica_count - max_replica_count. desiredReplicaCount: type: integer format: int32 description: >- The desired number of replicas for this deployment. This represents the target replica count that the system is trying to achieve. readOnly: true replicaCount: type: integer format: int32 readOnly: true autoscalingPolicy: $ref: "#/components/schemas/gatewayAutoscalingPolicy" baseModel: type: string title: >- The base model name. e.g. accounts/fireworks/models/falcon-7b acceleratorCount: type: integer format: int32 description: >- The number of accelerators used per replica. If not specified, the default is the estimated minimum required by the base model. acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: The type of accelerator to use. precision: $ref: "#/components/schemas/DeploymentPrecision" description: The precision with which the model should be served. maxConcurrencyPerReplica: type: integer format: int32 description: >- The maximum number of concurrent (in-flight) requests a single replica will accept before shedding load. Requests that arrive while a replica is already at this limit are rejected early with HTTP 429 instead of queueing — a per-replica admission gate for controlling tail latency. When unset (0), the platform default is used. cluster: type: string description: >- If set, this deployment is deployed to a cloud-premise cluster. readOnly: true enableAddons: type: boolean description: If true, PEFT addons are enabled for this deployment. draftTokenCount: type: integer format: int32 description: >- The number of candidate tokens to generate per step for speculative decoding. Default is the base model's draft_token_count. Set CreateDeploymentRequest.disable_speculative_decoding to false to disable this behavior. draftModel: type: string description: >- The draft model name for speculative decoding. e.g. accounts/fireworks/models/my-draft-model If empty, speculative decoding using a draft model is disabled. Default is the base model's default_draft_model. Set CreateDeploymentRequest.disable_speculative_decoding to false to disable this behavior. ngramSpeculationLength: type: integer format: int32 description: >- The length of previous input sequence to be considered for N-gram speculation. enableSessionAffinity: type: boolean description: |- Whether to apply sticky routing based on `user` field. Serverless will be set to true when creating deployment. directRouteApiKeys: type: array items: type: string description: >- The set of API keys used to access the direct route deployment. If direct routing is not enabled, this field is unused. numPeftDeviceCached: type: integer format: int32 title: How many peft adapters to keep on gpu side for caching directRouteType: $ref: "#/components/schemas/gatewayDirectRouteType" description: >- If set, this deployment will expose an endpoint that bypasses the Fireworks API gateway. directRouteHandle: type: string description: >- The handle for calling a direct route. The meaning of the handle depends on the direct route type of the deployment: INTERNET -> The host name for accessing the deployment GCP_PRIVATE_SERVICE_CONNECT -> The service attachment name used to create the PSC endpoint. AWS_PRIVATELINK -> The service name used to create the VPC endpoint. readOnly: true deploymentTemplate: type: string description: >- The name of the deployment template to use for this deployment. Only available to enterprise accounts. autoTune: $ref: "#/components/schemas/gatewayAutoTune" description: The performance profile to use for this deployment. placement: $ref: "#/components/schemas/gatewayPlacement" description: >- The desired geographic region where the deployment must be placed. If unspecified, the default is the GLOBAL multi-region. region: $ref: "#/components/schemas/gatewayRegion" description: >- The geographic region where the deployment is presently located. This region may change over time, but within the `placement` constraint. readOnly: true maxContextLength: type: integer format: int32 description: >- The maximum context length supported by the model (context window). If set to 0 or not specified, the model's default maximum context length will be used. updateTime: type: string format: date-time description: The update time for the deployment. readOnly: true disableDeploymentSizeValidation: type: boolean description: Whether the deployment size validation is disabled. enableHotLoad: type: boolean description: Whether to use hot load for this deployment. hotLoadBucketType: $ref: "#/components/schemas/DeploymentHotLoadBucketType" title: >- hot load bucket name, indicate what type of storage to use for hot load enableHotReloadLatestAddon: type: boolean description: >- Allows up to 1 addon at a time to be loaded, and will merge it into the base model. deploymentShape: type: string description: >- The name of the deployment shape that this deployment is using. On the server side, this will be replaced with the deployment shape version name. activeModelVersion: type: string description: >- The model version that is currently active and applied to running replicas of a deployment. targetModelVersion: type: string description: >- The target model version that is being rolled out to the deployment. In a ready steady state, the target model version is the same as the active model version. replicaStats: $ref: "#/components/schemas/gatewayReplicaStats" description: >- Per-replica deployment status counters. Provides visibility into the deployment process by tracking replicas in different stages of the deployment lifecycle. readOnly: true hotLoadBucketUrl: type: string title: >- For hot load bucket location e.g for s3: s3://mybucket/; for GCS: gs://mybucket/, no trailing slash pricingPlanId: type: string description: >- Optional pricing plan ID for custom billing configuration. If set, this deployment will use the pricing plan's billing rules instead of default billing behavior. hotLoadTrainerJob: type: string title: >- Trainer job whose hot_load_bucket_url this deployment should use. At creation, the trainer's bucket URL is copied into this deployment's hot_load_bucket_url. The deployment continues working after the trainer is deleted (snapshot-at-creation semantics). Format: accounts/{account}/rlorTrainerJobs/{job} hotLoadTransitionType: $ref: "#/components/schemas/DeploymentHotLoadTransitionType" description: >- Controls hot-load reload semantics for this deployment. ASYNC (default when enable_hot_load is set) pauses the generator/prefiller mid-flight and skips draining in-flight requests during hot load. SYNC drains in-flight requests before applying the new model. preemptible: type: boolean description: When true, this deployment runs as preemptible. title: >- The properties of the deployment being updated. `deployment.name` must be populated with the updated resource's name. required: - baseModel description: |- The properties of the deployment being updated. `deployment.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Deployment operationId: Gateway_DeleteDeployment responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: hard description: If true, this will perform a hard deletion. in: query required: false schema: type: boolean - name: ignoreChecks description: >- If true, this will ignore checks and force the deletion of a deployment that is currently deployed and is in use. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}/ledger: servers: - url: https://api.fireworks.ai get: summary: Get ledger operationId: Gateway_GetLedger responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetLedgerResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Reset ledger for hot load operationId: Gateway_ResetLedger responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}/shards: servers: - url: https://api.fireworks.ai get: summary: List Deployment Shards operationId: Gateway_ListDeploymentShards responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDeploymentShardsResponse" parameters: - name: pageSize description: >- The maximum number of deployment shards to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDeploymentShards call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDeploymentShards must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Only deployment shard satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "create_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}/terminationMessage: servers: - url: https://api.fireworks.ai get: summary: Get termination message from deployment operationId: Gateway_GetTerminationMessage responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetTerminationMessageResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}:metrics: servers: - url: https://api.fireworks.ai get: summary: Get Deployment Metrics (Deprecated) operationId: Gateway_GetDeploymentMetrics responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetDeploymentMetricsResponse" parameters: - name: timeRange description: >- The time range to fetch metrics for (e.g. "1m", "10m", "2h"). Defaults to 10m. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}:scale: servers: - url: https://api.fireworks.ai patch: summary: Scale Deployment to a specific number of replicas or to zero operationId: Gateway_ScaleDeployment responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayScaleDeploymentBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/deployments/{deployment_id}:undelete: servers: - url: https://api.fireworks.ai post: summary: Undelete Deployment operationId: Gateway_UndeleteDeployment responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeployment" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: deployment_id in: path required: true description: The Deployment Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayUndeleteDeploymentBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/developerPasses: servers: - url: https://api.fireworks.ai get: summary: List Developer Passes operationId: Gateway_ListDeveloperPasses responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDeveloperPassesResponse" parameters: - name: pageSize description: Maximum number of developer passes to return. in: query required: false schema: type: integer format: int32 - name: pageToken description: Page token from a previous ListDeveloperPasses call. in: query required: false schema: type: string - name: filter description: Filter expression (e.g., "state=ACTIVE") in: query required: false schema: type: string - name: orderBy description: Order by expression (e.g., "create_time desc") in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: |- CRUD APIs for developer passes. Create Developer Pass operationId: Gateway_CreateDeveloperPass responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeveloperPass" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayDeveloperPass" description: The developer pass to create. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/developerPasses/{developer_passe_id}: servers: - url: https://api.fireworks.ai get: summary: Get Developer Pass operationId: Gateway_GetDeveloperPass responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeveloperPass" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: developer_passe_id in: path required: true description: The Developer Passe Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Developer Pass operationId: Gateway_UpdateDeveloperPass responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDeveloperPass" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: developer_passe_id in: path required: true description: The Developer Passe Id schema: type: string requestBody: content: application/json: schema: type: object properties: autoRenew: type: boolean description: >- Whether the developer pass will automatically renew upon expiry. endTime: type: string format: date-time description: The time at which the developer pass ends. readOnly: true createTime: type: string format: date-time description: Creation timestamp. readOnly: true updateTime: type: string format: date-time description: Last update timestamp. readOnly: true state: $ref: "#/components/schemas/gatewayDeveloperPassState" title: State of the developer pass readOnly: true statusMessage: type: string description: >- Additional context for the current state, such as a payment decline reason. readOnly: true lastRenewTime: type: string format: date-time description: The time at which the developer pass was last renewed. readOnly: true version: $ref: "#/components/schemas/DeveloperPassVersion" description: The version of developer pass (V1 = weekly, V2 = monthly). readOnly: true annotations: $ref: "#/components/schemas/gatewayDeveloperPassAnnotation" readOnly: true title: The developer pass to update. description: The developer pass to update. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/dpoJobs: servers: - url: https://api.fireworks.ai get: operationId: Gateway_ListDpoJobs responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListDpoJobsResponse" parameters: - name: pageSize description: >- The maximum number of dpo jobs to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListDpoJobs call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListDpoJobs must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Filter criteria for the returned jobs. See https://google.aip.dev/160 for the filter syntax specification. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: includeArchived description: >- If true, include archived (post-delete retention) jobs in the response. Archived jobs are inert (no resources running, no further billing) but their checkpoints remain promotable until the retention window expires. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: operationId: Gateway_CreateDpoJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDpoJob" parameters: - name: dpoJobId description: ID of the DPO job, a random ID will be generated if not specified. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayDpoJob" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/dpoJobs/{dpo_job_id}: servers: - url: https://api.fireworks.ai get: operationId: Gateway_GetDpoJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDpoJob" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: dpo_job_id in: path required: true description: The Dpo Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: operationId: Gateway_DeleteDpoJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dpo_job_id in: path required: true description: The Dpo Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/dpoJobs/{dpo_job_id}:cancel: servers: - url: https://api.fireworks.ai post: summary: Cancel Dpo Job operationId: Gateway_CancelDpoJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dpo_job_id in: path required: true description: The Dpo Job Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCancelDpoJobBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/dpoJobs/{dpo_job_id}:getMetricsFileEndpoint: servers: - url: https://api.fireworks.ai get: operationId: Gateway_GetDpoJobMetricsFileEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetDpoJobMetricsFileResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dpo_job_id in: path required: true description: The Dpo Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/dpoJobs/{dpo_job_id}:resume: servers: - url: https://api.fireworks.ai post: summary: Resume Dpo Job operationId: Gateway_ResumeDpoJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayDpoJob" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: dpo_job_id in: path required: true description: The Dpo Job Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayResumeDpoJobBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluationJobs: servers: - url: https://api.fireworks.ai get: summary: List Evaluation Jobs operationId: Gateway_ListEvaluationJobs responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListEvaluationJobsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter in: query required: false schema: type: string - name: orderBy in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Evaluation Job operationId: Gateway_CreateEvaluationJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluationJob" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateEvaluationJobBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluationJobs/{evaluation_job_id}: servers: - url: https://api.fireworks.ai get: summary: Get Evaluation Job operationId: Gateway_GetEvaluationJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluationJob" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluation_job_id in: path required: true description: The Evaluation Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Evaluation Job operationId: Gateway_DeleteEvaluationJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluation_job_id in: path required: true description: The Evaluation Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluationJobs/{evaluation_job_id}:getExecutionLogEndpoint: servers: - url: https://api.fireworks.ai get: summary: Get Evaluation Job execution logs (stream log endpoint + tracing IDs). operationId: Gateway_GetEvaluationJobExecutionLogEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetEvaluationJobExecutionLogEndpointResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluation_job_id in: path required: true description: The Evaluation Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluations: servers: - url: https://api.fireworks.ai get: summary: List Evaluations operationId: Gateway_ListEvaluations responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListEvaluationsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter in: query required: false schema: type: string - name: orderBy in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Evaluation operationId: Gateway_CreateEvaluation responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluation" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateEvaluationBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluations/{evaluation_id}: servers: - url: https://api.fireworks.ai get: summary: Get Evaluation operationId: Gateway_GetEvaluation responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluation" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluation_id in: path required: true description: The Evaluation Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: |- TODO: Add Update Evaluation Update Evaluation rpc UpdateEvaluation(UpdateEvaluationRequest) returns (Evaluation) { option (google.api.http) = { patch: "/v1/{evaluation.name=accounts/*/evaluations/*}" body: "evaluation" }; } Delete Evaluation operationId: Gateway_DeleteEvaluation responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluation_id in: path required: true description: The Evaluation Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluations/{evaluation_id}:preview: servers: - url: https://api.fireworks.ai post: summary: Preview an evaluation with sample data operationId: Gateway_PreviewEvaluation responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPreviewEvaluationResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluation_id in: path required: true description: The Evaluation Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayPreviewEvaluationBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluations:validateAssertions: servers: - url: https://api.fireworks.ai post: summary: Validate evaluation assertions operationId: Gateway_ValidateAssertions responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayValidateAssertionsResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayValidateAssertionsBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators: servers: - url: https://api.fireworks.ai get: summary: List Evaluators description: Lists all evaluators for an account with pagination support. operationId: Gateway_ListEvaluators responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListEvaluatorsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter in: query required: false schema: type: string - name: orderBy in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Evaluator (Deprecated) description: "Deprecated: Use CreateEvaluatorV2 instead" operationId: Gateway_CreateEvaluator responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluator" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateEvaluatorBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}: servers: - url: https://api.fireworks.ai get: summary: Get Evaluator description: >- Retrieves an evaluator by name. Use this to monitor build progress after creation (**step 6** in the [Create Evaluator](/api-reference/create-evaluator) workflow). Possible states: - `BUILDING` - Environment is being prepared - `ACTIVE` - Evaluator is ready to use - `BUILD_FAILED` - Check build logs via [Get Evaluator Build Log Endpoint](/api-reference/get-evaluator-build-log-endpoint) operationId: Gateway_GetEvaluator responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluator" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Evaluator description: >- Updates evaluator metadata (display_name, description, default_dataset). Changing `requirements` or `entry_point` triggers a rebuild. To upload new source code, set `prepare_code_upload: true` then follow the upload flow. operationId: Gateway_UpdateEvaluator responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluator" parameters: - name: prepareCodeUpload description: >- If true, prepare a new code upload/build attempt by transitioning the evaluator to BUILDING state. Can be used without update_mask. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true updateTime: type: string format: date-time readOnly: true state: $ref: "#/components/schemas/gatewayEvaluatorState" readOnly: true criteria: type: array items: type: object $ref: "#/components/schemas/gatewayCriterion" title: >- Criteria for the evaluator, it should produce a score for the metric (name of criteria) Used for eval3 with UI upload path requirements: type: string title: Content for the requirements.txt for package installation entryPoint: type: string title: >- entry point of the evaluator inside the codebase. In module::function or path::function format status: $ref: "#/components/schemas/gatewayStatus" title: >- Status of the evaluator, used to expose build status to the user readOnly: true commitHash: type: string title: >- Commit hash of this evaluator from the user's original codebase source: $ref: "#/components/schemas/EvaluatorSource" description: Source information for the evaluator codebase. defaultDataset: type: string title: Default dataset that is associated with the evaluator encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true title: The evaluator to update. The name must be set. description: The evaluator to update. The name must be set. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Evaluator description: Deletes an evaluator and its associated versions and build artifacts. operationId: Gateway_DeleteEvaluator responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}/versions: servers: - url: https://api.fireworks.ai get: summary: List Evaluator Revisions operationId: Gateway_ListEvaluatorVersions responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListEvaluatorVersionsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter in: query required: false schema: type: string - name: orderBy description: >- Default order should be reverse chronological (newest first) per AIP-162. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}/versions/{version_id}: servers: - url: https://api.fireworks.ai get: summary: |- CRUD APIs for evaluator revisions (AIP-162). Get Evaluator Revision operationId: Gateway_GetEvaluatorVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluatorVersion" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: >- Delete Evaluator Revision (deletes alias if name is an alias per AIP-162) operationId: Gateway_DeleteEvaluatorVersion responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}/versions/{version_id}:alias: servers: - url: https://api.fireworks.ai post: summary: Alias Evaluator Revision - assign or update a user-specified alias operationId: Gateway_AliasEvaluatorVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluatorVersion" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayAliasEvaluatorVersionBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}/versions/{version_id}:rollback: servers: - url: https://api.fireworks.ai post: summary: Rollback Evaluator to a specific revision operationId: Gateway_RollbackEvaluator responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluatorVersion" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayRollbackEvaluatorBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}:getBuildLogEndpoint: servers: - url: https://api.fireworks.ai get: summary: Get Evaluator Build Log Endpoint description: |- Returns a signed URL to download the evaluator's build logs. Useful for debugging `BUILD_FAILED` state. operationId: Gateway_GetEvaluatorBuildLogEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetEvaluatorBuildLogEndpointResponse parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}:getSourceCodeSignedUrl: servers: - url: https://api.fireworks.ai get: summary: Get Evaluator Source Code Endpoint description: |- Returns a signed URL to download the evaluator's source code archive. Useful for debugging or reviewing the uploaded code. operationId: Gateway_GetEvaluatorSourceCodeEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetEvaluatorSourceCodeEndpointResponse parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}:getUploadEndpoint: servers: - url: https://api.fireworks.ai post: summary: Get Evaluator Upload Endpoint description: >- Returns signed URLs for uploading evaluator source code (**step 3** in the [Create Evaluator](/api-reference/create-evaluator) workflow). After receiving the signed URL, upload your `.tar.gz` archive using HTTP `PUT` with `Content-Type: application/octet-stream` header. operationId: Gateway_GetEvaluatorUploadEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetEvaluatorUploadEndpointResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayGetEvaluatorUploadEndpointBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators/{evaluator_id}:validateUpload: servers: - url: https://api.fireworks.ai post: summary: Validate Evaluator Upload description: >- Triggers server-side validation of the uploaded source code (**step 5** in the [Create Evaluator](/api-reference/create-evaluator) workflow). The server extracts and processes the archive, then builds the evaluator environment. Poll [Get Evaluator](/api-reference/get-evaluator) to monitor progress. operationId: Gateway_ValidateEvaluatorUpload responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: evaluator_id in: path required: true description: The Evaluator Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayValidateEvaluatorUploadBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluators:previewEvaluator: servers: - url: https://api.fireworks.ai post: operationId: Gateway_PreviewEvaluator responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPreviewEvaluatorResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayPreviewEvaluatorBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/evaluatorsV2: servers: - url: https://api.fireworks.ai post: summary: Create Evaluator description: >- Creates a custom evaluator for scoring model outputs. Evaluators use the [Eval Protocol](https://evalprotocol.io) to define test cases, run model inference, and score responses. They are used with evaluation jobs and Reinforcement Fine-Tuning (RFT). ## Source Code Requirements Your project should contain: - `requirements.txt` - Python dependencies for your evaluator - `test_*.py` - Pytest test file(s) with [`@evaluation_test`](https://evalprotocol.io/reference/evaluation-test) decorated functions - Any additional code/modules your evaluator needs ## Workflow **Recommended:** Use the [`ep upload`](https://evalprotocol.io/reference/cli#ep-upload) CLI command to handle all these steps automatically. If using the API directly: 1. Call this endpoint to create the evaluator resource 2. Package your source directory as a `.tar.gz` (respecting `.gitignore`) 3. Call [Get Evaluator Upload Endpoint](/api-reference/get-evaluator-upload-endpoint) to get a signed upload URL 4. `PUT` the tar.gz file to the signed URL 5. Call [Validate Evaluator Upload](/api-reference/validate-evaluator-upload) to trigger server-side validation 6. Poll [Get Evaluator](/api-reference/get-evaluator) until ready Once active, reference the evaluator in [Create Evaluation Job](/api-reference/create-evaluation-job) or [Create Reinforcement Fine-tuning Job](/api-reference/create-reinforcement-fine-tuning-job). operationId: Gateway_CreateEvaluatorV2 responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayEvaluator" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateEvaluatorV2Body" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/externalKeys: servers: - url: https://api.fireworks.ai get: summary: List External Keys operationId: Gateway_ListExternalKeys responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListExternalKeysResponse" parameters: - name: pageSize description: Page size. in: query required: false schema: type: integer format: int32 - name: pageToken description: Page token. in: query required: false schema: type: string - name: filter description: Filter expression. in: query required: false schema: type: string - name: orderBy description: Order by. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create External Key operationId: Gateway_CreateExternalKey responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayExternalKey" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayExternalKey" description: External key to create. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/externalKeys/{external_key_id}: servers: - url: https://api.fireworks.ai get: summary: Get External Key operationId: Gateway_GetExternalKey responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayExternalKey" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: external_key_id in: path required: true description: The External Key Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update External Key operationId: Gateway_UpdateExternalKey responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayExternalKey" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: external_key_id in: path required: true description: The External Key Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: Display name for the external key. createTime: type: string format: date-time description: Creation timestamp. readOnly: true updateTime: type: string format: date-time description: Last update timestamp. readOnly: true state: $ref: "#/components/schemas/gatewayExternalKeyState" description: Current state of the external key. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the external key status. readOnly: true title: External key to update. description: External key to update. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete External Key operationId: Gateway_DeleteExternalKey responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: external_key_id in: path required: true description: The External Key Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/identityProviders: servers: - url: https://api.fireworks.ai get: summary: List Identity Providers operationId: Gateway_ListIdentityProviders responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListIdentityProvidersResponse" parameters: - name: pageSize description: Page size in: query required: false schema: type: integer format: int32 - name: pageToken description: Page token in: query required: false schema: type: string - name: filter description: Filter expression in: query required: false schema: type: string - name: orderBy description: Order by in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Identity Provider operationId: Gateway_CreateIdentityProvider responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayIdentityProvider" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayIdentityProvider" description: Identity provider to create required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/identityProviders/{identity_provider_id}: servers: - url: https://api.fireworks.ai get: summary: Get Identity Provider operationId: Gateway_GetIdentityProvider responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayIdentityProvider" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: identity_provider_id in: path required: true description: The Identity Provider Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Identity Provider operationId: Gateway_UpdateIdentityProvider responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayIdentityProvider" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: identity_provider_id in: path required: true description: The Identity Provider Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string title: Display name for the identity provider createTime: type: string format: date-time title: Creation timestamp readOnly: true updateTime: type: string format: date-time title: Last update timestamp readOnly: true samlConfig: $ref: "#/components/schemas/gatewaySamlConfig" oidcConfig: $ref: "#/components/schemas/gatewayOidcConfig" tenantDomains: type: array items: type: string title: >- List of allowed domains for this identity provider Example: ["example.com", "example.co.uk", "example.de"] If not provided, domain will be derived from the account email state: $ref: "#/components/schemas/gatewayIdentityProviderState" title: Current state of the identity provider readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the identity provider status. readOnly: true domainUrl: type: string description: The domain URL. readOnly: true issuerUrl: type: string description: The OIDC issuer URL. readOnly: true clientId: type: string description: The OIDC client ID. readOnly: true enableJitUserProvisioning: type: boolean description: >- Enable Just-In-Time (JIT) user provisioning. When enabled, users are automatically created in the account on first SSO login if they don't already exist. When disabled, users must be pre-provisioned before they can authenticate via SSO. jitDefaultRole: type: string description: >- Default role assigned to JIT-provisioned users. Valid values: "admin", "user", "contributor", "inference-user". Only applies when enable_jit_user_provisioning is true and RBAC V2 is enabled. If empty or unset, defaults to "inference-user" (least privilege). If RBAC V2 is not enabled for the account, JIT users always get "user" role. enforceSso: type: boolean title: >- Enforce SSO authentication and restrict account access to users with approved email domains. When enabled: - Users with email domains matching tenant_domains must authenticate via this identity provider - Users with other email domains are blocked (tenant_domains becomes an allowlist) - Superusers and API keys are exempt from this enforcement enableIdpInitiatedSso: type: boolean description: >- Enable IdP-initiated SAML (Security Assertion Markup Language) single sign-on. When enabled, users can start the login flow from their identity provider's portal (e.g., Okta app launcher) instead of from the Fireworks login page. Only supported for SAML identity providers. title: Identity provider to update description: Identity provider to update required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Identity Provider operationId: Gateway_DeleteIdentityProvider responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: identity_provider_id in: path required: true description: The Identity Provider Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models: servers: - url: https://api.fireworks.ai get: summary: List Models operationId: Gateway_ListModels responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListModelsResponse" parameters: - name: pageSize description: >- The maximum number of models to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListModels call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListModels must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only model satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Model operationId: Gateway_CreateModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayModel" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateModelBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}: servers: - url: https://api.fireworks.ai get: summary: Get Model operationId: Gateway_GetModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayModel" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Model operationId: Gateway_UpdateModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayModel" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: |- Human-readable display name of the model. e.g. "My Model" Must be fewer than 64 characters long. description: type: string description: >- The description of the model. Must be fewer than 1000 characters long. createTime: type: string format: date-time description: The creation time of the model. readOnly: true state: $ref: "#/components/schemas/gatewayModelState" description: The state of the model. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: >- Contains detailed message when the last model operation fails. readOnly: true kind: $ref: "#/components/schemas/ModelKind" description: |- The kind of model. If not specified, the default is HF_PEFT_ADDON. githubUrl: type: string description: The URL to GitHub repository of the model. huggingFaceUrl: type: string description: The URL to the Hugging Face model. baseModelDetails: $ref: "#/components/schemas/gatewayBaseModelDetails" description: >- Base model details. Required if kind is HF_BASE_MODEL. Must not be set otherwise. peftDetails: $ref: "#/components/schemas/gatewayPEFTDetails" description: |- PEFT addon details. Required if kind is HF_PEFT_ADDON or HF_TEFT_ADDON. teftDetails: $ref: "#/components/schemas/gatewayTEFTDetails" description: >- TEFT addon details. Required if kind is HF_TEFT_ADDON. Must not be set otherwise. public: type: boolean description: If true, the model will be publicly readable. conversationConfig: $ref: "#/components/schemas/gatewayConversationConfig" description: >- If set, the Chat Completions API will be enabled for this model. contextLength: type: integer format: int32 description: The maximum context length supported by the model. supportsImageInput: type: boolean description: If set, images can be provided as input to the model. supportsTools: type: boolean description: >- If set, tools (i.e. functions) can be provided as input to the model, and the model may respond with one or more tool calls. importedFrom: type: string description: >- The name of the the model from which this was imported. This field is empty if the model was not imported. readOnly: true fineTuningJob: type: string description: >- If the model was created from a fine-tuning job, this is the fine-tuning job name. readOnly: true defaultDraftModel: type: string description: >- The default draft model to use when creating a deployment. If empty, speculative decoding is disabled by default. defaultDraftTokenCount: type: integer format: int32 description: >- The default draft token count to use when creating a deployment. Must be specified if default_draft_model is specified. deployedModelRefs: type: array items: type: object $ref: "#/components/schemas/gatewayDeployedModelRef" description: Populated from GetModel API call only. readOnly: true cluster: type: string description: >- The resource name of the BYOC cluster to which this model belongs. e.g. accounts/my-account/clusters/my-cluster. Empty if it belongs to a Fireworks cluster. readOnly: true deprecationDate: $ref: "#/components/schemas/typeDate" description: >- If specified, this is the date when the serverless deployment of the model will be taken down. calibrated: type: boolean description: >- If true, the model is calibrated and can be deployed to non-FP16 precisions. readOnly: true tunable: type: boolean description: >- Deprecated: V1 training stack only — LoRA only, limited architecture support. If the model has use_training_v2=true, use supervised_lora_tunable and supervised_full_parameter_tunable instead. readOnly: true supportsLora: type: boolean description: Whether this model supports LoRA. useHfApplyChatTemplate: type: boolean description: >- If true, the model will use the Hugging Face apply_chat_template API to apply the chat template. updateTime: type: string format: date-time description: The update time for the model. readOnly: true defaultSamplingParams: type: object additionalProperties: type: number format: float description: >- A json object that contains the default sampling parameters for the model. readOnly: true rlTunable: type: boolean description: >- Deprecated: V1 training stack only — LoRA only, limited architecture support. If the model has use_training_v2=true, use rl_lora_tunable and rl_full_parameter_tunable instead. readOnly: true trainingContextLength: type: integer format: int32 description: The maximum context length supported by the model. snapshotType: $ref: "#/components/schemas/ModelSnapshotType" supportsServerless: type: boolean description: If true, the model has a serverless deployment. readOnly: true useTrainingV2: type: boolean description: >- If true, SFT jobs for this base model use service-mode (StatefulSet + orchestration sidecar) instead of the legacy batch Job path. supervisedLoraTunable: type: boolean description: >- V2 only. Whether the model supports LoRA supervised fine-tuning and DPO (lora_rank > 0). True when a validated LORA_TRAINER training shape exists. readOnly: true supervisedFullParameterTunable: type: boolean description: >- V2 only. Whether the model supports full-parameter supervised fine-tuning and DPO (lora_rank = 0). True when a validated POLICY_TRAINER training shape exists. readOnly: true rlLoraTunable: type: boolean description: >- V2 only. Whether the model supports LoRA reinforcement learning (lora_rank > 0). True when a validated LORA_TRAINER training shape exists plus a deployment shape. readOnly: true rlFullParameterTunable: type: boolean description: >- V2 only. Whether the model supports full-parameter reinforcement learning (lora_rank = 0). True when validated POLICY_TRAINER + LORA_TRAINER training shapes exist plus a deployment shape. readOnly: true encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true serverlessModes: type: array items: type: object $ref: "#/components/schemas/gatewayServerlessMode" description: >- The serverless modes this base model is available on — the serverless read shape. Populated on read (e.g. the serverless catalog read path); one entry per serving mode. Empty for non-serverless models. readOnly: true title: |- The properties of the model being updated. `model.name` must be populated with the updated resource's name. description: |- The properties of the model being updated. `model.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Model operationId: Gateway_DeleteModel responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}/versions: servers: - url: https://api.fireworks.ai post: summary: |- CRUD APIs for model versions. Create Model Version operationId: Gateway_CreateModelVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayModelVersion" parameters: - name: versionId in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayModelVersion" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}/versions/{version_id}:deploy: servers: - url: https://api.fireworks.ai post: summary: Deploy Model Version operationId: Gateway_DeployModelVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/googlelongrunningOperation" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayDeployModelVersionBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}/versions:count: servers: - url: https://api.fireworks.ai get: summary: Returns the number of model versions. operationId: Gateway_GetModelVersionCount responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetModelVersionCountResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}:getDownloadEndpoint: servers: - url: https://api.fireworks.ai get: summary: Get Model Download Endpoint operationId: Gateway_GetModelDownloadEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetModelDownloadEndpointResponse" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}:getUploadEndpoint: servers: - url: https://api.fireworks.ai post: summary: Get Model Upload Endpoint operationId: Gateway_GetModelUploadEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayGetModelUploadEndpointResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayGetModelUploadEndpointBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}:import: servers: - url: https://api.fireworks.ai post: summary: Transfer model from S3 to GCP storage operationId: Gateway_ImportModel responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/googlelongrunningOperation" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayImportModelBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}:prepare: servers: - url: https://api.fireworks.ai post: summary: Prepare Model for different precisions operationId: Gateway_PrepareModel responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayPrepareModelBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/models/{model_id}:validateUpload: servers: - url: https://api.fireworks.ai get: summary: Validate Model Upload operationId: Gateway_ValidateModelUpload responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayValidateModelUploadResponse" parameters: - name: skipHfConfigValidation description: If true, skip the Hugging Face config validation. in: query required: false schema: type: boolean - name: trustRemoteCode description: If true, trusts remote code when validating the Hugging Face config. in: query required: false schema: type: boolean - name: configOnly description: If true, skip tokenizer and parameter name validation. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: model_id in: path required: true description: The Model Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/policySettings: servers: - url: https://api.fireworks.ai get: summary: Returns the singleton PolicySettings for the given account. operationId: Gateway_GetPolicySettings responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPolicySettings" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Updates the per-account PolicySettings singleton. operationId: Gateway_UpdatePolicySettings responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPolicySettings" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: type: object properties: defaultPermissions: $ref: "#/components/schemas/PolicySettingsModelPermissions" description: >- Per-account default permissions. If unset, defaults to allow-all. This powers the UI's "Default" row. rules: type: array items: type: object $ref: "#/components/schemas/PolicySettingsModelAccessRule" description: >- Per-model override rows. A model not listed here uses default_permissions. updateTime: type: string format: date-time readOnly: true cmekRequired: type: boolean description: >- Whether new customer-data resources for this account must be CMEK-encrypted. Readable by the account, but only a Fireworks superuser can change it (the write path gates this explicitly). Enabling requires a READY primary ExternalKey. title: >- The policy settings to update. `policy_settings.name` must be populated. Format: accounts/{account}/policySettings description: >- The policy settings to update. `policy_settings.name` must be populated. Format: accounts/{account}/policySettings required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/pricingPlans: servers: - url: https://api.fireworks.ai get: summary: List Pricing Plans operationId: Gateway_ListPricingPlans responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListPricingPlansResponse" parameters: - name: pageSize description: Maximum number of pricing plans to return in: query required: false schema: type: integer format: int32 - name: pageToken description: Page token from a previous ListPricingPlans call in: query required: false schema: type: string - name: filter description: Filter expression (e.g., "state=READY") in: query required: false schema: type: string - name: orderBy description: Order by expression (e.g., "create_time desc") in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/pricingPlans/{pricing_plan_id}: servers: - url: https://api.fireworks.ai get: summary: |- CRUD APIs for pricing plans. Get Pricing Plan operationId: Gateway_GetPricingPlan responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPricingPlan" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: pricing_plan_id in: path required: true description: The Pricing Plan Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/quotas: servers: - url: https://api.fireworks.ai get: summary: List Quotas description: Lists all quotas for an account. operationId: Gateway_ListQuotas responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListQuotasResponse" parameters: - name: pageSize description: >- The maximum number of quotas to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListQuotas call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListQuotas must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only quota satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/quotas/{quota_id}: servers: - url: https://api.fireworks.ai get: summary: Get Quota description: Gets a single quota by resource name. operationId: Gateway_GetQuota responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayQuota" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: quota_id in: path required: true description: The Quota Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Quota description: Updates a quota. operationId: Gateway_UpdateQuota responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayQuota" parameters: - name: allowMissing description: If true, and the quota does not exist, it will be created. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string - name: quota_id in: path required: true description: The Quota Id schema: type: string requestBody: content: application/json: schema: type: object properties: value: type: string format: int64 description: >- The value of the quota being enforced. This may be lower than the max_value if the user manually lowers it. maxValue: type: string format: int64 description: The maximum approved value. usage: type: number format: double description: The usage of the quota. readOnly: true updateTime: type: string format: date-time description: The update time for the quota. readOnly: true title: |- The properties of the quota being updated. `quota.name` must be populated with the updated resource's name. description: |- The properties of the quota being updated. `quota.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/reinforcementFineTuningJobs: servers: - url: https://api.fireworks.ai get: summary: List Reinforcement Fine-tuning Jobs operationId: Gateway_ListReinforcementFineTuningJobs responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListReinforcementFineTuningJobsResponse parameters: - name: pageSize description: >- The maximum number of fine-tuning jobs to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListReinforcementLearningFineTuningJobs call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListReinforcementLearningFineTuningJobs must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Filter criteria for the returned jobs. See https://google.aip.dev/160 for the filter syntax specification. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Reinforcement Fine-tuning Job operationId: Gateway_CreateReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayReinforcementFineTuningJob" parameters: - name: reinforcementFineTuningJobId description: >- ID of the reinforcement fine-tuning job, a random UUID will be generated if not specified. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayReinforcementFineTuningJob" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/reinforcementFineTuningJobs/{reinforcement_fine_tuning_job_id}: servers: - url: https://api.fireworks.ai get: summary: Get Reinforcement Fine-tuning Job operationId: Gateway_GetReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayReinforcementFineTuningJob" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Reinforcement Fine-tuning Job operationId: Gateway_UpdateReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayReinforcementFineTuningJob" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string createTime: type: string format: date-time readOnly: true completedTime: type: string format: date-time description: The completed time for the reinforcement fine-tuning job. readOnly: true dataset: type: string description: The name of the dataset used for training. evaluationDataset: type: string description: The name of a separate dataset to use for evaluation. evalAutoCarveout: type: boolean description: Whether to auto-carve the dataset for eval. state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true createdBy: type: string description: >- The email address of the user who initiated this fine-tuning job. readOnly: true trainingConfig: $ref: "#/components/schemas/gatewayBaseTrainingConfig" description: Common training configurations. evaluator: type: string description: The evaluator resource name to use for RLOR fine-tuning job. wandbConfig: $ref: "#/components/schemas/gatewayWandbConfig" description: >- The Weights & Biases team/user account for logging training progress. awsS3Config: $ref: "#/components/schemas/gatewayAwsS3Config" description: The AWS configuration for S3 dataset access. azureBlobStorageConfig: $ref: "#/components/schemas/gatewayAzureBlobStorageConfig" description: The Azure configuration for Blob Storage dataset access. outputStats: type: string description: >- The output dataset's aggregated stats for the evaluation job. readOnly: true jobProgress: $ref: "#/components/schemas/gatewayJobProgress" description: Job progress. readOnly: true inferenceParameters: $ref: >- #/components/schemas/gatewayReinforcementFineTuningJobInferenceParameters description: RFT inference parameters. chunkSize: type: integer format: int32 description: >- Data chunking for rollout, default size 200, enabled when dataset > 300. Valid range is 1-10,000. outputMetrics: type: string readOnly: true maxInferenceReplicaCount: type: integer format: int32 title: |- Maximum number of replicas to use for the deployment. Default is 1 nodeCount: type: integer format: int32 description: |- The number of nodes to use for the fine-tuning job. If not specified, the default is 1. lossConfig: $ref: "#/components/schemas/gatewayReinforcementLearningLossConfig" description: >- Reinforcement learning loss method + hyperparameters for the underlying trainers. trainerLogsSignedUrl: type: string description: >- The signed URL for the trainer logs file (stdout/stderr). Only populated if the account has trainer log reading enabled. readOnly: true acceleratorSeconds: type: object additionalProperties: type: string format: int64 description: >- Accelerator seconds used by the job, keyed by accelerator type (e.g., "NVIDIA_H100_80GB"). Updated when job completes or is cancelled. readOnly: true maxConcurrentRollouts: type: integer format: int32 description: Maximum number of concurrent rollouts during the RFT job. maxConcurrentEvaluations: type: integer format: int32 description: Maximum number of concurrent evaluations during the RFT job. purpose: $ref: "#/components/schemas/gatewayPurpose" description: Scheduling purpose for this job. encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true title: The reinforcement fine-tuning job to update. required: - dataset - evaluator description: The reinforcement fine-tuning job to update. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Reinforcement Fine-tuning Job operationId: Gateway_DeleteReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/reinforcementFineTuningJobs/{reinforcement_fine_tuning_job_id}:cancel: servers: - url: https://api.fireworks.ai post: summary: Cancel Reinforcement Fine-tuning Job operationId: Gateway_CancelReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayCancelReinforcementFineTuningJobBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/reinforcementFineTuningJobs/{reinforcement_fine_tuning_job_id}:debug: servers: - url: https://api.fireworks.ai post: summary: Debug Reinforcement Fine-tuning Job operationId: Gateway_DebugReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayDebugReinforcementFineTuningJobResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayDebugReinforcementFineTuningJobBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/reinforcementFineTuningJobs/{reinforcement_fine_tuning_job_id}:getMetricsFileEndpoint: servers: - url: https://api.fireworks.ai get: operationId: Gateway_GetReinforcementFineTuningJobMetricsFileEndpoint responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetReinforcementFineTuningJobMetricsFileResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/reinforcementFineTuningJobs/{reinforcement_fine_tuning_job_id}:resume: servers: - url: https://api.fireworks.ai post: summary: Resume Reinforcement Fine-tuning Job operationId: Gateway_ResumeReinforcementFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayReinforcementFineTuningJob" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: reinforcement_fine_tuning_job_id in: path required: true description: The Reinforcement Fine-tuning Job Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayResumeReinforcementFineTuningJobBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs: servers: - url: https://api.fireworks.ai get: summary: List Reinforcement Fine-tuning Steps operationId: Gateway_ListRlorTrainerJobs responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListRlorTrainerJobsResponse" parameters: - name: pageSize description: >- The maximum number of fine-tuning jobs to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListRlorTuningJobs call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListRlorTuningJobs must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Filter criteria for the returned jobs. See https://google.aip.dev/160 for the filter syntax specification. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: includeArchived description: >- If true, include archived (post-delete retention) jobs in the response. Archived jobs are inert (no resources running, no further billing) but their checkpoints remain promotable until the retention window expires. Default false matches the customer-facing list semantics — archived jobs are hidden unless explicitly requested. in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Reinforcement Fine-tuning Step operationId: Gateway_CreateRlorTrainerJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRlorTrainerJob" parameters: - name: rlorTrainerJobId description: >- ID of the RLOR trainer job, a random UUID will be generated if not specified. in: query required: false schema: type: string - name: trainingShape description: >- Optional validated training-shape selector for service-mode launches. Accepted formats: - accounts/{account}/trainingShapes/{shape} - accounts/{account}/trainingShapes/{shape}/versions/{version} - accounts/{account}/trainingShapes/{shape}/versions/latest When a shape (without /versions/*) is provided, the latest validated version is used. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayRlorTrainerJob" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs/{rlor_trainer_job_id}: servers: - url: https://api.fireworks.ai get: summary: Get Reinforcement Fine-tuning Step operationId: Gateway_GetRlorTrainerJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRlorTrainerJob" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Reinforcement Fine-tuning Step operationId: Gateway_DeleteRlorTrainerJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs/{rlor_trainer_job_id}/checkpoints: servers: - url: https://api.fireworks.ai get: summary: List checkpoints for an RLOR Trainer Job operationId: Gateway_ListRlorTrainerJobCheckpoints responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListRlorTrainerJobCheckpointsResponse parameters: - name: pageSize description: Maximum number of checkpoints to return. Default is 50, max is 200. in: query required: false schema: type: integer format: int32 - name: pageToken description: Page token from a previous call. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs/{rlor_trainer_job_id}/checkpoints/{checkpoint_id}:sync: servers: - url: https://api.fireworks.ai post: summary: |- Ensure an RLOR training checkpoint is present in the requested storage availability class. By default this copies a single-region runtime checkpoint into managed durable storage. Set request.target_availability=SINGLE_REGION to stage a durable checkpoint into request.region before restoring there. operationId: Gateway_SyncRlorTrainerJobCheckpoint responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewaySyncRlorTrainerJobCheckpointResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string - name: checkpoint_id in: path required: true description: The Checkpoint Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewaySyncRlorTrainerJobCheckpointBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs/{rlor_trainer_job_id}/publicLogs: servers: - url: https://api.fireworks.ai get: summary: Get a signed URL for the public trainer logs of an RLOR Trainer Job. operationId: Gateway_GetRlorTrainerJobPublicLogs responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetRlorTrainerJobPublicLogsResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs/{rlor_trainer_job_id}:cancel: servers: - url: https://api.fireworks.ai post: summary: Cancel Rlor Trainer Job operationId: Gateway_CancelRlorTrainerJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCancelRlorTrainerJobBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/rlorTrainerJobs/{rlor_trainer_job_id}:resume: servers: - url: https://api.fireworks.ai post: summary: Resume Rlor Trainer Job operationId: Gateway_ResumeRlorTrainerJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRlorTrainerJob" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: rlor_trainer_job_id in: path required: true description: The Rlor Trainer Job Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayResumeRlorTrainerJobBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/routers: servers: - url: https://api.fireworks.ai get: summary: List Routers operationId: Gateway_ListRouters responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListRoutersResponse" parameters: - name: pageSize description: >- The maximum number of routers to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListRouters call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListRouters must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Filter criteria for the returned routers. See https://google.aip.dev/160 for the filter syntax specification. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Router operationId: Gateway_CreateRouter responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRouter" parameters: - name: routerId description: ID of the router. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayRouter" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/routers/{router_id}: servers: - url: https://api.fireworks.ai get: summary: |- CRUD APIs for routers. Get Router operationId: Gateway_GetRouter responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRouter" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: router_id in: path required: true description: The Router Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Router operationId: Gateway_UpdateRouter responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayRouter" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: router_id in: path required: true description: The Router Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true state: $ref: "#/components/schemas/gatewayRouterState" description: The state of the router. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true deployments: type: array items: type: string description: The deployment names to be covered by the router. model: type: string description: |- The model name to route requests to. model is only applicable to single-region deployments. For multi-region deployments, model must be empty. weightedRandom: $ref: "#/components/schemas/gatewayWeightedRandomStrategy" evenLoad: $ref: "#/components/schemas/gatewayEvenLoadStrategy" aliases: type: array items: type: string title: >- Aliases for the router. These are the alias names that can be used to route requests through the router. By default, these aliases are generated by the system. Examples: "accounts/foo/deployments/my-deployment" or "foo/my-deployment" "accounts/foo/routers/my-router" "accounts/foo/models/my-base-model" - only if this is the first deployment using this model readOnly: true autoGenerated: type: boolean description: >- auto_generated indicates this router was generated by the system for a deployment. Auto-generated routers cannot be updated or deleted by the user. readOnly: true public: type: boolean description: >- True if the router is public (any account can query the underlying workload), false if the router is private (only the account that owns the router can query the underlying workload). title: |- The properties of the router being updated. `router.name` must be populated with the updated resource's name. description: |- The properties of the router being updated. `router.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Router operationId: Gateway_DeleteRouter responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: router_id in: path required: true description: The Router Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/secrets: servers: - url: https://api.fireworks.ai get: summary: List Secrets description: >- Lists all secrets for an account. Note that the `value` field is not returned in the response for security reasons. Only the `name` and `key_name` fields are included for each secret. operationId: Gateway_ListSecrets responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListSecretsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter description: Unused but required to use existing ListRequest functionality. in: query required: false schema: type: string - name: orderBy description: Unused but required to use existing ListRequest functionality. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: operationId: Gateway_CreateSecret responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySecret" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewaySecret" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/secrets/{secret_id}: servers: - url: https://api.fireworks.ai get: summary: Get Secret description: >- Retrieves a secret by name. Note that the `value` field is not returned in the response for security reasons. Only the `name` and `key_name` fields are included. operationId: Gateway_GetSecret responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySecret" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: secret_id in: path required: true description: The Secret Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: operationId: Gateway_UpdateSecret responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySecret" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: secret_id in: path required: true description: The Secret Id schema: type: string requestBody: content: application/json: schema: type: object properties: keyName: type: string title: >- name of the key. In this case, it can be WOLFRAM_ALPHA_API_KEY value: type: string example: sk-1234567890abcdef description: >- The secret value. This field is INPUT_ONLY and will not be returned in GET or LIST responses for security reasons. The value is only accepted when creating or updating secrets. required: - keyName required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: operationId: Gateway_DeleteSecret responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: secret_id in: path required: true description: The Secret Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/serverlessModels: servers: - url: https://api.fireworks.ai get: summary: List serverless models available to the account (public catalog). operationId: Gateway_ListServerlessModels responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListServerlessModelsResponse" parameters: - name: pageSize description: >- The maximum number of models to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: A page token, received from a previous ListServerlessModels call. in: query required: false schema: type: string - name: filter description: AIP-160 filter. See https://google.aip.dev/160 for the grammar. in: query required: false schema: type: string - name: orderBy description: A comma-separated list of fields to order by. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/serverlessRateLimits: servers: - url: https://api.fireworks.ai get: summary: >- Lists effective global rate limits for shared serverless pool deployments (fireworks-owned). Values reflect configured limits as observed in production monitoring. operationId: Gateway_ListAccountServerlessRateLimits responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListAccountServerlessRateLimitsResponse parameters: - name: deployment description: If set, only return limits for this deployment resource name. in: query required: false schema: type: string - name: start description: >- Start of the time range for time series data. Defaults to 30 days before `end`. in: query required: false schema: type: string format: date-time - name: end description: >- End of the time range for time series data. Defaults to the request time. in: query required: false schema: type: string format: date-time - name: interval description: >- Step size for the time series: each point is the peak effective limit observed within that window. Defaults to 4 hours. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/serverlessTokenUsage: servers: - url: https://api.fireworks.ai get: summary: >- Serverless pool token throughput (tokens per minute) by time bucket and base model, from production usage metrics. operationId: Gateway_GetAccountServerlessTokenUsage responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayGetAccountServerlessTokenUsageResponse parameters: - name: start description: Start of the time range. Defaults to 30 days before `end`. in: query required: false schema: type: string format: date-time - name: end description: End of the time range. Defaults to the request time. in: query required: false schema: type: string format: date-time - name: interval description: >- Step size for each point (peak or average TPM within each window). Defaults to 4 hours. in: query required: false schema: type: string - name: includePeakTokensPerMinuteByBaseModel description: >- Whether to include each section in the response. At least one must be true; otherwise the request fails with INVALID_ARGUMENT. in: query required: false schema: type: boolean - name: includeAverageTokensPerMinuteByBaseModel in: query required: false schema: type: boolean - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/supervisedFineTuningJobs: servers: - url: https://api.fireworks.ai get: summary: List Supervised Fine-tuning Jobs operationId: Gateway_ListSupervisedFineTuningJobs responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListSupervisedFineTuningJobsResponse parameters: - name: pageSize description: >- The maximum number of fine-tuning jobs to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListSupervisedFineTuningJobs call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListSupervisedFineTuningJobs must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Filter criteria for the returned jobs. See https://google.aip.dev/160 for the filter syntax specification. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Supervised Fine-tuning Job operationId: Gateway_CreateSupervisedFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySupervisedFineTuningJob" parameters: - name: supervisedFineTuningJobId description: >- ID of the supervised fine-tuning job, a random UUID will be generated if not specified. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewaySupervisedFineTuningJob" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/supervisedFineTuningJobs/{supervised_fine_tuning_job_id}: servers: - url: https://api.fireworks.ai get: summary: Get Supervised Fine-tuning Job operationId: Gateway_GetSupervisedFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySupervisedFineTuningJob" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: supervised_fine_tuning_job_id in: path required: true description: The Supervised Fine-tuning Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Supervised Fine-tuning Job operationId: Gateway_DeleteSupervisedFineTuningJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: supervised_fine_tuning_job_id in: path required: true description: The Supervised Fine-tuning Job Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/supervisedFineTuningJobs/{supervised_fine_tuning_job_id}:cancel: servers: - url: https://api.fireworks.ai post: summary: Cancel Supervised Fine-tuning Job operationId: Gateway_CancelSupervisedFineTuningJob responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: supervised_fine_tuning_job_id in: path required: true description: The Supervised Fine-tuning Job Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayCancelSupervisedFineTuningJobBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/supervisedFineTuningJobs/{supervised_fine_tuning_job_id}:resume: servers: - url: https://api.fireworks.ai post: summary: Resume Supervised Fine-tuning Job operationId: Gateway_ResumeSupervisedFineTuningJob responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewaySupervisedFineTuningJob" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: supervised_fine_tuning_job_id in: path required: true description: The Supervised Fine-tuning Job Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayResumeSupervisedFineTuningJobBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/supervisedFineTuningJobs:estimateCost: servers: - url: https://api.fireworks.ai post: summary: Estimate the cost of a Supervised Fine-tuning Job operationId: Gateway_EstimateSupervisedFineTuningJobCost responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayEstimateSupervisedFineTuningJobCostResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewaySupervisedFineTuningJob" description: >- uses the dataset, base model, and number of epochs to determine the estimated cost of the fine-tuning job. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trails: servers: - url: https://api.fireworks.ai get: summary: List Trails operationId: Gateway_ListTrails responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListTrailsResponse" parameters: - name: pageSize description: >- The maximum number of trails to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListTrails call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListTrails must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: >- Filter criteria for the returned trails. See https://google.aip.dev/160 for the filter syntax specification. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "create_time,display_name" The default sort order is ascending. To specify descending order for a field, append a " desc" suffix. e.g. "create_time desc" If not specified, the default order is by "create_time desc". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create Trail operationId: Gateway_CreateTrail responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrail" parameters: - name: trailId description: >- Optional ID for the trail. If not specified, a random UUID will be generated. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayTrail" description: The trail to create required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trails/{trail_id}: servers: - url: https://api.fireworks.ai get: summary: |- CRUD APIs for trails. Get Trail operationId: Gateway_GetTrail responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrail" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: trail_id in: path required: true description: The Trail Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Trail operationId: Gateway_UpdateTrail responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrail" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: trail_id in: path required: true description: The Trail Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string title: Human-readable display name for the trail createTime: type: string format: date-time title: Creation timestamp readOnly: true updateTime: type: string format: date-time title: Last update timestamp readOnly: true description: type: string title: Optional description of what this trail is tracking createdBy: type: string title: The email address of the user who created this trail readOnly: true langfuseConfig: type: string description: >- Langfuse configuration for this trail, inherited from the account at creation time. Contains API credentials for a Project within the customer's Langfuse Organization. Serialized JSON format: {"public_key":"pk-lf-xxx","secret_key":"sk-lf-xxx","host":"https://langfuse-prod.fireworks.ai"} If empty, the default Langfuse Organization/Project will be used. readOnly: true defaultModel: type: string description: >- Default model for requests through this trail. Format: "{provider}/{model_id}" (e.g., "anthropic/claude-3-5-sonnet-20240620"). Can be overridden per request. providerKey: type: string description: >- Provider API key for this trail. When creating a trail: provide the raw API key (e.g., "sk-ant-api03-xxxx...") After creation: this field contains a secret reference (e.g., "accounts/{account_id}/secrets/trail-xxx-provider-key") The LiteLLM gateway retrieves the actual key from Secret Manager using this reference. Can be overridden by specifying api_key in the request body. title: |- The properties of the trail being updated. `trail.name` must be populated with the updated resource's name. description: |- The properties of the trail being updated. `trail.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Trail operationId: Gateway_DeleteTrail responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: trail_id in: path required: true description: The Trail Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trails/{trail_id}:getMetrics: servers: - url: https://api.fireworks.ai get: summary: Get Trail Metrics (trace counts, token usage from tracing service) operationId: Gateway_GetTrailMetrics responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrailMetrics" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: trail_id in: path required: true description: The Trail Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingRuns: servers: - url: https://api.fireworks.ai get: summary: >- List training runs under an account (Tinker RestClient list_training_runs). operationId: Gateway_ListTrainingRuns responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListTrainingRunsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter in: query required: false schema: type: string - name: orderBy in: query required: false schema: type: string - name: readMask in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: >- Create a training run: the durable, account-owned model lineage for a Tinker create_lora_training_client. A session may host many runs, and a run may span sessions across resume. operationId: Gateway_CreateTrainingRun responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingRun" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingRun" description: |- The run to create. Carries base_model, lora_rank, and the source_training_session_id the run is created under. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingRuns/{training_run_id}: servers: - url: https://api.fireworks.ai get: summary: Get a training run (Tinker RestClient get_training_run parity). operationId: Gateway_GetTrainingRun responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingRun" parameters: - name: readMask in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_run_id in: path required: true description: The Training Run Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingRuns/{training_run_id}:abortMaterialization: servers: - url: https://api.fireworks.ai post: summary: >- Abort a pending training run when trainer materialization failed or timed out. INTERNAL: api_gateway calls this on failed Tinker create_model. operationId: Gateway_AbortTrainingRunMaterialization responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingRun" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_run_id in: path required: true description: The Training Run Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayAbortTrainingRunMaterializationBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingRuns/{training_run_id}:finalizeMaterialization: servers: - url: https://api.fireworks.ai post: summary: |- Finalize a training run after the trainer has successfully materialized its run-scoped model id. INTERNAL: api_gateway calls this after proxying Tinker create_model to the trainer. operationId: Gateway_FinalizeTrainingRunMaterialization responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingRun" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_run_id in: path required: true description: The Training Run Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayFinalizeTrainingRunMaterializationBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions: servers: - url: https://api.fireworks.ai get: summary: List training sessions under an account. operationId: Gateway_ListTrainingSessions responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListTrainingSessionsResponse" parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter in: query required: false schema: type: string - name: orderBy in: query required: false schema: type: string - name: readMask in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create a training session under an account. operationId: Gateway_CreateTrainingSession responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingSession" parameters: - name: trainingSessionId in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingSession" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}: servers: - url: https://api.fireworks.ai get: summary: Get a training session. operationId: Gateway_GetTrainingSession responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingSession" parameters: - name: readMask in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: |- Delete a training session. Releases the shared-trainer binding before removing the row; the session transitions DELETING → gone. operationId: Gateway_DeleteTrainingSession responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}/checkpoints: servers: - url: https://api.fireworks.ai get: summary: |- List the checkpoints saved during a TrainingSession. Access is scoped to the session's account. operationId: Gateway_ListTrainingSessionCheckpoints responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListTrainingSessionCheckpointsResponse parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken description: |- Opaque cursor from a previous response's next_page_token; resume listing after it. Do not parse or construct this value. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}/checkpoints/{checkpoint_id}:promote: servers: - url: https://api.fireworks.ai post: summary: |- Promote a checkpoint saved during a TrainingSession into a customer-owned Model, identified by checkpoint_name. Only promotable (inference) checkpoints can be promoted. operationId: Gateway_PromoteTrainingSessionCheckpoint responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayPromoteTrainingSessionCheckpointResponse parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string - name: checkpoint_id in: path required: true description: The Checkpoint Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayPromoteTrainingSessionCheckpointBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}/chunkOutputs: servers: - url: https://api.fireworks.ai get: summary: List chunk completion events on a TrainingSession. operationId: Gateway_ListTrainingSessionChunkOutputs responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListTrainingSessionChunkOutputsResponse parameters: - name: pageSize in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter description: |- AIP filter expression evaluated against TrainingSessionChunkOutput's FilterDefs (step, end_of_epoch, epoch_id, output_dataset_name, etc). Empty = no filter. in: query required: false schema: type: string - name: orderBy description: >- AIP order-by clause evaluated against TrainingSessionChunkOutput's OrderDefs. Empty defaults to "step asc" (chunks appear in training order). in: query required: false schema: type: string - name: readMask in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: |- Create a chunk-completion row on a TrainingSession (per-chunk distributions, dataset refs, checkpoints). Trainer-driven write; the server validates the parent session and materializes the row. operationId: Gateway_CreateTrainingSessionChunkOutput responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingSessionChunkOutput" parameters: - name: trainingSessionChunkOutputId description: >- Optional client-chosen id; server generates one if empty. The natural key on the row is (session, step) — server rejects duplicates. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingSessionChunkOutput" description: |- Chunk payload. The trainer supplies step, distribution snapshot, dataset refs, and checkpoint pointers; the server fills in `name`, `create_time`, `update_time` and signs URL fields on read. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}/trainingCurves: servers: - url: https://api.fireworks.ai get: summary: List scalar training curves for line charts on a TrainingSession. operationId: Gateway_ListTrainingSessionTrainingCurves responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListTrainingSessionTrainingCurvesResponse parameters: - name: metricKeys description: |- Metric keys to project (a row-selector over the returned curves, NOT an AIP-160 predicate — see `filter` below). Empty = all metrics. in: query required: false explode: true schema: type: array items: type: string - name: maxPoints description: >- Optional server-side downsampling cap per curve. 0 disables downsampling. in: query required: false schema: type: integer format: int32 - name: pageSize description: Standard List fields (AIP-132). in: query required: false schema: type: integer format: int32 - name: pageToken in: query required: false schema: type: string - name: filter description: >- AIP-160 filter — predicate over curve rows (e.g. step ranges). Distinct from `metric_keys`, which is a projection over which metrics to return. in: query required: false schema: type: string - name: orderBy in: query required: false schema: type: string - name: readMask in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}:loadState: servers: - url: https://api.fireworks.ai post: summary: Load or switch the LoRA adapter state for a training session. operationId: Gateway_LoadTrainingSessionState responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingSession" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayLoadTrainingSessionStateBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingSessions/{training_session_id}:logMetrics: servers: - url: https://api.fireworks.ai post: summary: >- Log scalar training metrics on a TrainingSession (append-only time-series, upserts on (session, key, step) so retries are idempotent). operationId: Gateway_LogTrainingSessionMetrics responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayLogTrainingSessionMetricsResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_session_id in: path required: true description: The Training Session Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayLogTrainingSessionMetricsBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingShapes: servers: - url: https://api.fireworks.ai get: summary: List Training Shapes operationId: Gateway_ListTrainingShapes responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListTrainingShapesResponse" parameters: - name: pageSize description: >- The maximum number of resources to return. Max page_size is 200; values above 200 are coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: A page token from a previous ListTrainingShapes call. in: query required: false schema: type: string - name: filter description: Filter per AIP-160. in: query required: false schema: type: string - name: orderBy description: Order by fields, default "create_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: |- CRUD APIs for training shape. Create Training Shape operationId: Gateway_CreateTrainingShape responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingShape" parameters: - name: trainingShapeId description: >- The ID of the training shape. If not specified, a random ID will be generated. Must follow AIP-122 segment format and start with a letter. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingShape" description: The properties of the training shape being created. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingShapes/{training_shape_id}: servers: - url: https://api.fireworks.ai get: summary: Get Training Shape operationId: Gateway_GetTrainingShape responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingShape" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_shape_id in: path required: true description: The Training Shape Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Training Shape operationId: Gateway_UpdateTrainingShape responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingShape" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_shape_id in: path required: true description: The Training Shape Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: >- Human-readable display name of the training shape. e.g. "Llama3 70B H200 BF16" Must be fewer than 64 characters long. description: type: string description: >- The description of the training shape. Must be fewer than 1000 characters long. createTime: type: string format: date-time description: The creation time of the training shape. readOnly: true updateTime: type: string format: date-time description: The update time for the training shape. readOnly: true baseModel: type: string title: >- The base model name. e.g. accounts/fireworks/models/llama-3-70b deploymentShapeVersion: type: string title: >- The validated deployment shape version used for numerics verification. e.g. accounts/fireworks/deploymentShapes/rft-qwen3-4b/versions/abcd1234 trainerImageTag: type: string description: >- The validated trainer runtime image tag used for numerics verification. trainerMode: $ref: "#/components/schemas/TrainingShapeTrainerMode" description: Trainer mode used for the validated launch profile. nodeCount: type: integer format: int32 description: Node count validated for the launch profile. trainerShardingScheme: $ref: "#/components/schemas/gatewayTrainerShardingScheme" description: >- Structured sharding/parallelism profile validated for the trainer launch. modelType: type: string description: The model type of the base model (e.g. "llama", "qwen"). readOnly: true parameterCount: type: string format: int64 description: >- The parameter count of the base model, used for billion-parameter bucket matching. readOnly: true acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: The type of accelerator to use. acceleratorCount: type: integer format: int32 description: Total number of accelerators used by the job. baseModelWeightPrecision: $ref: "#/components/schemas/gatewayWeightPrecision" description: >- Default precision for base weights during training (prefer BF16; QLoRA/QAT omitted). maxSupportedContextLength: type: integer format: int32 description: Capacity limits validated for this shape. minTotalAcceleratorCount: type: integer format: int32 description: >- Minimum total accelerator count (world size, across all data-parallel replicas) this shape has been validated at. Written from shape-validation CI evidence; do not hand-edit downward. title: >- The properties of the training shape being updated; `training_shape.name` must be populated. required: - baseModel - trainerImageTag description: >- The properties of the training shape being updated; `training_shape.name` must be populated. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] delete: summary: Delete Training Shape operationId: Gateway_DeleteTrainingShape responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_shape_id in: path required: true description: The Training Shape Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingShapes/{training_shape_id}/versions: servers: - url: https://api.fireworks.ai get: summary: List Training Shapes Versions operationId: Gateway_ListTrainingShapeVersions responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListTrainingShapeVersionsResponse" parameters: - name: pageSize description: >- The maximum number of versions to return. Max page_size is 200; values above 200 are coerced. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: A page token from a previous ListTrainingShapeVersions call. in: query required: false schema: type: string - name: filter description: Filter per AIP-160. in: query required: false schema: type: string - name: orderBy description: Order by fields, default "create_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_shape_id in: path required: true description: The Training Shape Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/trainingShapes/{training_shape_id}/versions/{version_id}: servers: - url: https://api.fireworks.ai get: summary: Get Training Shape Version operationId: Gateway_GetTrainingShapeVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingShapeVersion" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_shape_id in: path required: true description: The Training Shape Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update Training Shape Version operationId: Gateway_UpdateTrainingShapeVersion responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayTrainingShapeVersion" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: training_shape_id in: path required: true description: The Training Shape Id schema: type: string - name: version_id in: path required: true description: The Version Id schema: type: string requestBody: content: application/json: schema: type: object properties: createTime: type: string format: date-time title: Creation time (used for ordering desc by default) readOnly: true snapshot: $ref: "#/components/schemas/gatewayTrainingShape" description: Full snapshot of the TrainingShape at this version. readOnly: true validated: type: boolean description: >- Whether this version has been validated through capacity tests. Only superusers can set this flag. public: type: boolean description: If true, this version will be publicly readable. latestValidated: type: boolean description: >- If true, this version is the latest validated version (at most one per shape). readOnly: true updateTime: type: string format: date-time description: Last update time of mutable fields such as validated/public. readOnly: true title: >- The properties of the version being updated; `training_shape_version.name` must be populated. description: >- The properties of the version being updated; `training_shape_version.name` must be populated. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/usageLimitIncreaseRequests: servers: - url: https://api.fireworks.ai get: summary: >- Lists the account's limit-increase requests (admin), filterable by state. operationId: Gateway_ListUserUsageLimitIncreaseRequests responses: "200": description: A successful response. content: application/json: schema: $ref: >- #/components/schemas/gatewayListUserUsageLimitIncreaseRequestsResponse parameters: - name: filter description: Filter expression. If empty, all requests are returned. in: query required: false schema: type: string - name: pageSize description: >- Maximum number of requests to return. Coerced to a default when unset and to a maximum when too large. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token from a previous call; only marks where the next page starts. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. Defaults to "create_time". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields are returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Files a member's request to raise their own per-user serverless cap. operationId: Gateway_CreateUserUsageLimitIncreaseRequest responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUserUsageLimitIncreaseRequest" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayUserUsageLimitIncreaseRequest" description: The request to create (user_id + requested_limit). required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/usageLimitIncreaseRequests/{usage_limit_increase_request_id}:approve: servers: - url: https://api.fireworks.ai post: summary: Approves a pending limit-increase request (admin). operationId: Gateway_ApproveUserUsageLimitIncreaseRequest responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUserUsageLimitIncreaseRequest" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: usage_limit_increase_request_id in: path required: true description: The Usage Limit Increase Request Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayApproveUserUsageLimitIncreaseRequestBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/usageLimitIncreaseRequests/{usage_limit_increase_request_id}:reject: servers: - url: https://api.fireworks.ai post: summary: Rejects a pending limit-increase request (admin). operationId: Gateway_RejectUserUsageLimitIncreaseRequest responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUserUsageLimitIncreaseRequest" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: usage_limit_increase_request_id in: path required: true description: The Usage Limit Increase Request Id schema: type: string requestBody: content: application/json: schema: $ref: >- #/components/schemas/gatewayGatewayRejectUserUsageLimitIncreaseRequestBody required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/usageLimits: servers: - url: https://api.fireworks.ai get: summary: |- Returns the account-level usage-limits singleton (default per-user cap, enforcement, account-wide usage rollup). operationId: Gateway_GetAccountUsageLimits responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayAccountUsageLimits" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: |- Updates the account-level usage-limits singleton (default per-user cap, enforcement). operationId: Gateway_UpdateAccountUsageLimits responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayAccountUsageLimits" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: type: object properties: defaultUserLimit: $ref: "#/components/schemas/typeMoney" description: >- The default per-user cap applied to every user without an override. Unset means no default cap (users are unlimited unless individually capped). enforcementMode: $ref: >- #/components/schemas/gatewayAccountUsageLimitsEnforcementMode description: >- Enforcement mode applied to the per-user caps in this account. used: $ref: "#/components/schemas/typeMoney" description: >- Account-wide current-period usage (the sum across users). Computed at read time; not stored. readOnly: true updateTime: type: string format: date-time description: >- The time the account usage limits (default cap / enforcement mode) were last updated. Unset until the account row is first written. readOnly: true title: >- The account usage limits to update. `account_usage_limits.name` must be set. description: >- The account usage limits to update. `account_usage_limits.name` must be set. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/userUsageLimits: servers: - url: https://api.fireworks.ai get: summary: |- Lists every user's usage limits in the account (the roster: usage vs. effective cap and over-limit status per user). operationId: Gateway_ListUserUsageLimits responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListUserUsageLimitsResponse" parameters: - name: pageSize description: >- Maximum number of users to return. Coerced to a default when unset and to a maximum when too large. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token from a previous ListUserUsageLimits call. page_size still bounds every page; the token only marks where the next page starts. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/users: servers: - url: https://api.fireworks.ai get: summary: List Users operationId: Gateway_ListUsers responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListUsersResponse" parameters: - name: pageSize description: |- The maximum number of users to return. The maximum page_size is 200, values above 200 will be coerced to 200. If unspecified, the default is 50. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- A page token, received from a previous ListUsers call. Provide this to retrieve the subsequent page. When paginating, all other parameters provided to ListUsers must match the call that provided the page token. in: query required: false schema: type: string - name: filter description: |- Only users satisfying the provided filter (if specified) will be returned. See https://google.aip.dev/160 for the filter grammar. in: query required: false schema: type: string - name: orderBy description: >- A comma-separated list of fields to order by. e.g. "foo,bar" The default sort order is ascending. To specify a descending order for a field, append a " desc" suffix. e.g. "foo desc,bar" Subfields are specified with a "." character. e.g. "foo.bar" If not specified, the default order is by "name". in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create User operationId: Gateway_CreateUser responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUser" parameters: - name: userId description: |- The user ID to use in the user name. e.g. my-user If not specified, a default ID is generated from user.email. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayUser" description: The properties of the user being created. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/users/{user_id}: servers: - url: https://api.fireworks.ai get: summary: Get User operationId: Gateway_GetUser responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUser" parameters: - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update User operationId: Gateway_UpdateUser responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUser" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: |- Human-readable display name of the user. e.g. "Alice" Must be fewer than 64 characters long. serviceAccount: type: boolean title: >- Whether this user is a service account (can only be set by admins) createTime: type: string format: date-time description: The creation time of the user. readOnly: true role: type: string description: >- The user's role: admin, user, contributor, inference-user, or custom. When set to "custom", the user's permissions are governed by permission_preset. email: type: string description: The user's email address. state: $ref: "#/components/schemas/gatewayUserState" description: The state of the user. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the user status. readOnly: true updateTime: type: string format: date-time description: The update time for the user. readOnly: true permissionPreset: type: string description: >- The permission preset for this user. Only valid when role is "custom". title: |- The properties of the User being updated. `user.name` must be populated with the updated resource's name. required: - role description: |- The properties of the User being updated. `user.name` must be populated with the updated resource's name. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/users/{user_id}/apiKeys: servers: - url: https://api.fireworks.ai get: summary: List API Keys operationId: Gateway_ListApiKeys responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayListApiKeysResponse" parameters: - name: pageSize description: >- Number of API keys to return in the response. Pagination support to be added. in: query required: false schema: type: integer format: int32 - name: pageToken description: >- Token for fetching the next page of results. Pagination support to be added. in: query required: false schema: type: string - name: filter description: Field for filtering results. in: query required: false schema: type: string - name: orderBy description: Field for ordering results. in: query required: false schema: type: string - name: readMask description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. in: query required: false schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] post: summary: Create API Key operationId: Gateway_CreateApiKey responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayApiKey" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayCreateApiKeyBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/users/{user_id}/apiKeys/accounts/{account_id}/users/{user_id}: servers: - url: https://api.fireworks.ai get: summary: Get API Key operationId: Gateway_GetApiKey responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayApiKey" parameters: - name: keyId description: The key ID for the API key. in: path required: true schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: Update API Key. Only `display_name` and `expire_time` may be modified; operationId: Gateway_UpdateApiKey responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayApiKey" parameters: - name: apiKey.keyId description: >- Unique identifier (Key ID) for the API key, used primarily for deletion. in: path required: true schema: type: string - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string requestBody: content: application/json: schema: type: object properties: displayName: type: string description: >- Display name for the API key, defaults to "default" if not specified. key: type: string description: >- The actual API key value, only available upon creation and not stored thereafter. readOnly: true createTime: type: string format: date-time description: Timestamp indicating when the API key was created. readOnly: true secure: type: boolean description: >- Indicates whether the plaintext value of the API key is unknown to Fireworks. If true, Fireworks does not know this API key's plaintext value. If false, Fireworks does know the plaintext value. readOnly: true email: type: string description: Email of the user who owns this API key. readOnly: true prefix: type: string title: >- The first few characters of the API key to visually identify it readOnly: true expireTime: type: string format: date-time description: >- Timestamp indicating when the API key will expire. If not set, the key never expires. annotations: type: object additionalProperties: type: string description: Optional key-value annotations for this API key. lastUsed: type: string format: date-time description: >- Timestamp indicating when the API key was last used. Not set if the key has never been used. readOnly: true isFirepass: type: boolean description: True when this key is a FirePass inference key. title: >- The API key to update. The `key_id` field is required to identify the target key; only fields named in `update_mask` are read from the rest of the message. description: >- The API key to update. The `key_id` field is required to identify the target key; only fields named in `update_mask` are read from the rest of the message. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/users/{user_id}/apiKeys:delete: servers: - url: https://api.fireworks.ai post: summary: Delete API Key operationId: Gateway_DeleteApiKey responses: "200": description: A successful response. content: application/json: schema: type: object properties: {} parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayDeleteApiKeyBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/users/{user_id}/usageLimits: servers: - url: https://api.fireworks.ai get: summary: |- Returns a single user's usage-limits singleton (override, effective cap, usage, and whether they are over limit). operationId: Gateway_GetUserUsageLimits responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUserUsageLimits" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string tags: - gateway.openapi_Gateway security: - BearerAuth: [] patch: summary: >- Updates a single user's usage-limits singleton (limit_override). Clearing the override reverts the user to the account default. operationId: Gateway_UpdateUserUsageLimits responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayUserUsageLimits" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string - name: user_id in: path required: true description: The User Id schema: type: string requestBody: content: application/json: schema: type: object properties: limitOverride: $ref: "#/components/schemas/typeMoney" description: >- The user's override cap. Unset means inherit the account default. effectiveLimit: $ref: "#/components/schemas/typeMoney" description: >- The user's resolved effective cap: their override, else the account default. Unset means the user has no cap (unlimited). Computed at read time; not stored. readOnly: true used: $ref: "#/components/schemas/typeMoney" description: Current-period usage. Computed at read time; not stored. readOnly: true enforcementMode: $ref: >- #/components/schemas/gatewayAccountUsageLimitsEnforcementMode description: >- Enforcement mode from the governing account policy. Read from the account usage-limits at read time; not stored on the user row. readOnly: true exceededUntil: type: string format: date-time description: >- When set, the user is over their limit and blocked until this time (the end of the current consumption period). Unset when the user is not over limit. Computed at read time; not stored. readOnly: true updateTime: type: string format: date-time description: >- The time this user's override was last set or changed. Unset when the user has no override (they inherit the account default); it tracks the override row, not account-default changes. readOnly: true title: >- The user usage limits to update. `user_usage_limits.name` must be set. Only limit_override is mutable; clearing it reverts the user to the account default. description: >- The user usage limits to update. `user_usage_limits.name` must be set. Only limit_override is mutable; clearing it reverts the user to the account default. required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}:testeval: servers: - url: https://api.fireworks.ai post: summary: >- Similar to preview evaluation, but no need to create the evaluation entry first. operationId: Gateway_TestEvaluation responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/gatewayPreviewEvaluationResponse" parameters: - name: account_id in: path required: true description: The Account Id schema: type: string requestBody: content: application/json: schema: $ref: "#/components/schemas/gatewayGatewayTestEvaluationBody" required: true tags: - gateway.openapi_Gateway security: - BearerAuth: [] /v1/accounts/{account_id}/datasets/{dataset_id}:upload: servers: - url: https://api.fireworks.ai post: summary: Upload Dataset Files description: > Provides a streamlined way to upload a dataset file in a single API request. This path can handle file sizes up to 150Mb. For larger file sizes use [Get Dataset Upload Endpoint](get-dataset-upload-endpoint). operationId: Gateway_UploadDatasetFile responses: "200": description: A successful response. content: application/json: schema: $ref: "#/components/schemas/FileUploadResponse" "400": description: Bad Request "401": description: Unauthorized "404": description: Not Found "500": description: Internal Server Error parameters: - name: account_id in: path required: true description: The account id schema: type: string - name: dataset_id in: path required: true description: The dataset id schema: type: string requestBody: content: multipart/form-data: schema: type: object properties: file: type: string format: binary required: true tags: - gateway-extra.openapi_Gateway security: - BearerAuth: [] /v1/responses: servers: - url: https://api.fireworks.ai/inference post: summary: Create Response description: >- Creates a model response, optionally interacting with custom tools via the Model Context Protocol (MCP). This endpoint supports conversational continuation and streaming. Explore our cookbooks for detailed examples: - [Basic MCP Usage](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/fireworks_mcp_examples.ipynb) - [Streaming with MCP](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/fireworks_mcp_with_streaming.ipynb) - [Conversational History with `previous_response_id`](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/fireworks_previous_response_cookbook.ipynb) - [Basic Streaming](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/fireworks_streaming_example.ipynb) - [Controlling Response Storage](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/mcp_server_with_store_false_argument.ipynb) operationId: create_response_v1_responses_post security: - BearerAuth: [] requestBody: required: true content: application/json: schema: $ref: "#/components/schemas/CreateResponse" responses: "200": description: Successful Response content: application/json: schema: $ref: "#/components/schemas/Response" "422": description: Validation Error content: application/json: schema: $ref: "#/components/schemas/HTTPValidationError" tags: - responses-harmonized.openapi_other get: summary: List Responses description: |- Get a list of all responses for the authenticated account. Args: limit: Maximum number of responses to return (default: 20, max: 100) after: Cursor for pagination - return responses after this ID before: Cursor for pagination - return responses before this ID operationId: list_responses_v1_responses_get security: - BearerAuth: [] parameters: - name: limit in: query required: false schema: type: integer default: 20 title: Limit - name: after in: query required: false schema: anyOf: - type: string - type: "null" title: After - name: before in: query required: false schema: anyOf: - type: string - type: "null" title: Before responses: "200": description: Successful Response content: application/json: schema: $ref: "#/components/schemas/ResponseList" "422": description: Validation Error content: application/json: schema: $ref: "#/components/schemas/HTTPValidationError" tags: - responses-harmonized.openapi_other /v1/responses/{response_id}: servers: - url: https://api.fireworks.ai/inference get: summary: Get Response operationId: get_response_v1_responses__response_id__get security: - BearerAuth: [] parameters: - name: response_id in: path required: true schema: type: string title: Response Id responses: "200": description: Successful Response content: application/json: schema: $ref: "#/components/schemas/Response" "422": description: Validation Error content: application/json: schema: $ref: "#/components/schemas/HTTPValidationError" tags: - responses-harmonized.openapi_other delete: summary: Delete Response description: >- Deletes a model response by its ID. Once deleted, the response data will be gone immediately and permanently. The response cannot be recovered and any conversations that reference this response ID will no longer be able to access it. operationId: delete_response_v1_responses__response_id__delete security: - BearerAuth: [] parameters: - name: response_id in: path required: true schema: type: string description: The ID of the response to delete title: Response Id description: The ID of the response to delete responses: "200": description: Successful Response content: application/json: schema: $ref: "#/components/schemas/DeleteResponse" "422": description: Validation Error content: application/json: schema: $ref: "#/components/schemas/HTTPValidationError" tags: - responses-harmonized.openapi_other /v1/completions: servers: - url: https://api.fireworks.ai/inference post: summary: Create Completion description: >- Create a completion for the provided prompt and parameters. For RL / agent rollouts, Fireworks inference exposes additional rollout-specific features: [`x-session-affinity` and `x-multi-turn-session-id`](https://docs.fireworks.ai/guides/rollout-inference#session-affinity) for multi-turn trajectories, and [MoE Router Replay (R3)](https://docs.fireworks.ai/guides/rollout-inference#moe-router-replay) for MoE expert tracing during rollouts. operationId: create_completion_v1_completions_post requestBody: content: application/json: schema: $ref: "#/components/schemas/CompletionRequest" required: true responses: "200": description: Successful Response content: application/json: schema: $ref: "#/components/schemas/CompletionResponse" text/event-stream: schema: $ref: "#/components/schemas/CompletionStreamResponse" "422": description: Validation Error content: application/json: schema: $ref: "#/components/schemas/HTTPValidationError" tags: - text-completion.openapi_other security: - BearerAuth: [] /v1/chat/completions: servers: - url: https://api.fireworks.ai/inference post: summary: Create Chat Completion description: >- Create a completion for the provided prompt and parameters. For RL / agent rollouts, Fireworks inference exposes additional rollout-specific features: [`x-session-affinity` and `x-multi-turn-session-id`](https://docs.fireworks.ai/guides/rollout-inference#session-affinity) for multi-turn trajectories, and [MoE Router Replay (R3)](https://docs.fireworks.ai/guides/rollout-inference#moe-router-replay) for MoE expert tracing during rollouts. operationId: create_chat_completion_v1_chat_completions_post requestBody: content: application/json: schema: $ref: "#/components/schemas/ChatCompletionRequest" required: true responses: "200": description: Successful Response content: application/json: schema: $ref: "#/components/schemas/ChatCompletionResponse" text/event-stream: schema: $ref: "#/components/schemas/ChatCompletionStreamResponse" "422": description: Validation Error content: application/json: schema: $ref: "#/components/schemas/HTTPValidationError" tags: - text-completion.openapi_other security: - BearerAuth: [] /v1/messages: servers: - url: https://api.fireworks.ai/inference post: summary: Create a Message description: >- **Anthropic-compatible endpoint.** Send a structured list of input messages with text and/or image content, and the model will generate the next message in the conversation. The Messages API can be used for either single queries or stateless multi-turn conversations. **Fireworks Quickstarts:** - [Serverless Quickstart](/getting-started/quickstart) - [Deployments Quickstart](/getting-started/ondemand-quickstart) operationId: messages_post responses: "200": description: Message object. content: application/json: schema: $ref: "#/components/schemas/AnthropicMessage" 4XX: description: >- Error response. See the [errors documentation](/guides/inference-error-codes) for more details. content: application/json: schema: $ref: "#/components/schemas/AnthropicErrorResponse" requestBody: content: application/json: schema: $ref: "#/components/schemas/AnthropicCreateMessageParams" required: true tags: - anthropic-messages.openapi_other security: - BearerAuth: [] components: securitySchemes: BearerAuth: type: http scheme: bearer description: >- Bearer authentication using your Fireworks API key. Format: Bearer bearerFormat: API_KEY schemas: AccountAccountType: type: string enum: - ACCOUNT_TYPE_UNSPECIFIED - ENTERPRISE default: ACCOUNT_TYPE_UNSPECIFIED AccountRateLimitMetric: type: string enum: - METRIC_UNSPECIFIED - TOKENS_GENERATED - TOKENS_PROMPT - TOKENS_CACHE_ADJUSTED_PROMPT - TOKENS_CACHED_PROMPT - TOKENS_UNCACHED_PROMPT default: METRIC_UNSPECIFIED description: Which rate-limit metric this row describes. AccountSuspendState: type: string enum: - UNSUSPENDED - FAILED_PAYMENTS - CREDIT_DEPLETED - MONTHLY_SPEND_LIMIT_EXCEEDED - BLOCKED_BY_ABUSE_RULE default: UNSUSPENDED AccountUsageDedicatedDeploymentUsage: type: object properties: deploymentId: type: string title: The deployment ID acceleratorType: type: string title: GPU type / accelerator type acceleratorSeconds: type: string format: int64 title: Accelerator seconds startTime: type: string format: date-time description: >- Start timestamp of the usage. Note: Dedicated deployment usage events are aggregated daily. Only the date portion (YYYY-MM-DD) is used; the time portion is ignored. endTime: type: string format: date-time description: >- End timestamp of the usage. Note: Dedicated deployment usage events are aggregated daily. Only the date portion (YYYY-MM-DD) is used; the time portion is ignored. baseModel: type: string title: Base model of the deployment usageType: type: string title: Usage type to distinguish between different deployment types placement: type: string description: >- Deployment placement as a string (e.g. US_IOWA_1, GLOBAL, US, EUROPE). Set to REGION_UNSPECIFIED when unknown. group: type: object additionalProperties: type: string description: >- Group-by dimension key-values (deployment_name, accelerator_type, team, project, environment). description: One dedicated deployment usage aggregation bucket. AccountUsageServerlessUsage: type: object properties: modelName: type: string description: 'Deprecated: use group["model_name"] instead.' promptTokens: type: string format: int64 title: Number of prompt tokens (for text inference) completionTokens: type: string format: int64 title: Number of completion tokens (for text inference) cachedPromptTokens: type: string format: int64 description: >- Number of prompt tokens served from cache (for text inference). Subset of prompt_tokens. uncachedPromptTokens: type: string format: int64 description: >- Number of prompt tokens not served from cache (for text inference). prompt_tokens - cached_prompt_tokens. startTime: type: string format: date-time title: >- Start timestamp of the usage. Note: Serverless usage events are aggregated daily. Only the date portion (YYYY-MM-DD) is used; endTime: type: string format: date-time title: >- End timestamp of the usage. Note: Serverless usage events are aggregated daily. Only the date portion (YYYY-MM-DD) is used; audioInputSeconds: type: number format: double title: Audio input seconds (for audio inference) usageType: type: string title: Usage type to distinguish between different inference types apiKeyId: type: string description: 'Deprecated: use group["api_key_id"] instead.' group: type: object additionalProperties: type: string description: >- Group-by dimension key-values (model_name, api_key_id, api_key_name, user_id, session_id, team, project, environment). session_id is present only for buckets that had an opaque client session; no-session usage omits the key (all such usage collapses into a single "no session" bucket). costNanoUsd: type: number format: double description: >- Cost in nano-USD (1e-9 USD), summed from properties.cost. 0 when absent (not "free"). Only huggingface currently stamps authoritative cost. description: One serverless usage aggregation bucket. AccountUsageTrainingUsage: type: object properties: jobId: type: string description: >- The managed training job id (e.g. supervised_fine_tuning_job id, rlor_trainer_job id). Empty for serverless training token rows; use `training_session_id` for those. Best-effort: extracted from billing event_id for legacy records that did not emit it as an explicit property. trainingSessionId: type: string description: >- The serverless training session id. Populated for SERVERLESS_TRAINING_TOKENS rows and empty for managed training jobs. jobType: type: string description: |- High-level training job kind: "sft", "dpo", "rft", "trainer", etc. Best-effort for legacy records. usageType: type: string description: >- The underlying billing usage_type as emitted to the billing pipeline, e.g. "SUPERVISED_FINE_TUNING_USAGE", "REINFORCEMENT_FINE_TUNING_GPU_SECONDS", "FULL_PARAM_FINE_TUNING_GPU_HOURS". Frontends can use this to pick the right metric to display (tokens vs. accelerator_seconds). tokens: type: string format: int64 description: |- Total tokens billed for this bucket. Populated for SFT / DPO usage; 0 for accelerator-second usage types. acceleratorType: type: string description: >- Accelerator type for accelerator-second usage (e.g. "NVIDIA_H100_80GB"). Empty for token-based usage. acceleratorSeconds: type: string format: int64 description: >- Total accelerator seconds billed for this bucket. Frontends typically display this as GPU hours (seconds / 3600). 0 for token-based usage. baseModel: type: string description: Base model for the training job, when known. startTime: type: string format: date-time description: >- Start timestamp of the usage bucket. Note: Training usage events are aggregated daily. Only the date portion (YYYY-MM-DD) is used; the time portion is ignored. endTime: type: string format: date-time description: >- End timestamp of the usage bucket. Note: Training usage events are aggregated daily. Only the date portion (YYYY-MM-DD) is used; the time portion is ignored. group: type: object additionalProperties: type: string description: >- Group-by dimension key-values for the requested group_by. May include any of: job_id, training_session_id, job_type, usage_type, accelerator_type, base_model, annotations.team, annotations.project, annotations.environment. Empty values are omitted. Annotation values will typically be empty for training usage until billing emitters propagate annotations. annotations: type: object additionalProperties: type: string description: |- Annotation key-values associated with this bucket (team, project, environment). Mirrors the annotation entries in `group` for API symmetry with future per-stream consumers. Typically empty for training usage today until trainer-side billing emitters populate properties.annotation_* on every record. description: >- One training usage aggregation bucket. Training usage is emitted by the trainer-side billing pipeline. SFT / DPO usage is metered in tokens; RFT and full-param (service-mode) trainer usage is metered in accelerator-seconds. Both metrics are returned on the same message; consumers should use `usage_type` to disambiguate which metric is meaningful for a given row. Some legacy training billing events did not populate every metadata field. Missing values are returned as the proto default (empty string / 0) rather than as an error. AssertionAssertionType: type: string enum: - ASSERTION_TYPE_UNSPECIFIED - ASSERTION_TYPE_LLM - ASSERTION_TYPE_CODE default: ASSERTION_TYPE_UNSPECIFIED BaseModelDetailsCheckpointFormat: type: string enum: - CHECKPOINT_FORMAT_UNSPECIFIED - NATIVE - HUGGINGFACE - UNINITIALIZED default: CHECKPOINT_FORMAT_UNSPECIFIED BatchInferenceJobLifecycleTimestamps: type: object properties: validatedTime: type: string format: date-time description: >- When dataset validation completed and the job left VALIDATING. Unset for jobs that skip validation or that fail before validating. readOnly: true runStartTime: type: string format: date-time description: >- When the runner first started processing (the job first became RUNNING). readOnly: true endTime: type: string format: date-time description: >- The terminal time of the job (when it reached COMPLETED, FAILED, EXPIRED, or CANCELLED). readOnly: true description: >- Observed execution milestone timestamps for the job, grouped into one sub-message so the job's lifecycle reads as a single cohesive unit rather than a scatter of top-level time fields. All OUTPUT_ONLY. CheckpointAvailabilityAvailabilityClass: type: string enum: - AVAILABILITY_CLASS_UNSPECIFIED - SINGLE_REGION - MANAGED_DURABLE default: AVAILABILITY_CLASS_UNSPECIFIED description: >- - SINGLE_REGION: Stored in one Fireworks runtime region. Optimized for low-latency hotload and reattach, but not highly available across regions. - MANAGED_DURABLE: Stored in Fireworks-managed durable storage for restore/export workflows. CodeAssertionExecutionOptions: type: object properties: timeoutMs: type: integer format: int32 title: Timeout in milliseconds, max 5 minutes memoryLimitMb: type: integer format: int32 title: Memory limit in MB, max 1GB envVars: type: object additionalProperties: type: string title: Environment variables title: Options for execution DatasetFormat: type: string enum: - FORMAT_UNSPECIFIED - CHAT - COMPLETION - RL default: FORMAT_UNSPECIFIED DeploymentHotLoadBucketType: type: string enum: - BUCKET_TYPE_UNSPECIFIED - MINIO - S3 - NEBIUS - FW_HOSTED default: BUCKET_TYPE_UNSPECIFIED title: "- FW_HOSTED: Fireworks hosted bucket" DeploymentHotLoadTransitionType: type: string enum: - HOT_LOAD_TRANSITION_TYPE_UNSPECIFIED - ASYNC - SYNC default: HOT_LOAD_TRANSITION_TYPE_UNSPECIFIED DeploymentPrecision: type: string enum: - PRECISION_UNSPECIFIED - FP16 - FP8 - FP8_MM - FP8_AR - FP8_MM_KV_ATTN - FP8_KV - FP8_MM_V2 - FP8_V2 - FP8_MM_KV_ATTN_V2 - NF4 - FP4 - BF16 - FP4_BLOCKSCALED_MM - FP4_MX_MOE default: PRECISION_UNSPECIFIED title: >- - PRECISION_UNSPECIFIED: if left unspecified we will treat this as a legacy model created before self serve DeploymentShapePresetType: type: string enum: - PRESET_TYPE_UNSPECIFIED - MINIMAL - FAST - THROUGHPUT - FULL_PRECISION - AGENTIC_CODING - CHAT - SUMMARIZATION - MULTI_LORA default: PRESET_TYPE_UNSPECIFIED title: |- - MINIMAL: Preset for cheapest & most minimal type of deployment - FAST: Preset for fastest generation & TTFT deployment - THROUGHPUT: Preset for best throughput deployment - FULL_PRECISION: Preset for deployment with full precision for training & most accurate numerics - AGENTIC_CODING: Preset for autonomous code generation and analysis for development workflows - CHAT: Preset for interactive conversational AI for customer engagement - SUMMARIZATION: Preset for efficient document and content summarization - MULTI_LORA: Preset for multi-LORA serving (deployments that can enable LORA addons) DeploymentShapeVersionCapability: type: string enum: - CAPABILITY_UNSPECIFIED - MULTI_LORA default: CAPABILITY_UNSPECIFIED description: A capability that a deployment shape version supports. DeveloperPassVersion: type: string enum: - VERSION_UNSPECIFIED - VERSION_V1 - VERSION_V2 default: VERSION_UNSPECIFIED EagleTrainingJobTrainingStyle: type: string enum: - TRAINING_STYLE_UNSPECIFIED - EAGLE - EAGLE3 - DFLASH - DSPARK default: TRAINING_STYLE_UNSPECIFIED EvaluatorSource: type: object properties: type: $ref: "#/components/schemas/EvaluatorSourceType" description: Identifies how the evaluator source code is provided. githubRepositoryName: type: string description: >- Normalized GitHub repository name (e.g. owner/repository) when the source is GitHub. EvaluatorSourceType: type: string enum: - TYPE_UNSPECIFIED - TYPE_UPLOAD - TYPE_GITHUB - TYPE_TEMPORARY default: TYPE_UNSPECIFIED title: |- - TYPE_UPLOAD: Source code is uploaded by the user - TYPE_GITHUB: Source code is from a GitHub repository - TYPE_TEMPORARY: Source code is a temporary UI uploaded code GetAccountUsageRequestFilterValues: type: object properties: values: type: array items: type: string description: >- Allowed values for one filter dimension; multiple values are OR'ed; different keys in `filter` are AND'ed. GetBillingSummaryRequestGranularity: type: string enum: - GRANULARITY_UNSPECIFIED - DAILY default: GRANULARITY_UNSPECIFIED description: |- Granularity of the per-bucket breakdown returned in usage_buckets. UNSPECIFIED returns only the range-level totals (no breakdown). GetReservationDataResponseReservationDataByType: type: object properties: acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" title: The accelerator type consumed: type: string format: int64 title: Number of GPUs currently consumed (in use) reserved: type: string format: int64 title: Number of GPUs reserved (total capacity) title: Reservation data for a specific accelerator type ListCostsResponseCostDataItem: type: object properties: subtotal: $ref: "#/components/schemas/typeMoney" description: The amount before any credits and discounts are applied. total: $ref: "#/components/schemas/typeMoney" description: The total amount after any credits and discounts have been applied. startTime: type: string format: date-time description: Costs returned are inclusive of `start_time`. endTime: type: string format: date-time description: Costs returned are exclusive of `end_time`. usageTotal: $ref: "#/components/schemas/typeMoney" description: >- Usage-based costs included in `total`, excluding fixed subscription fees. fixedTotal: $ref: "#/components/schemas/typeMoney" description: Fixed subscription fees included in `total`. ListPaymentMethodsResponseCard: type: object properties: brand: type: string description: Credit card brand. last4: type: string description: Last 4 digits of the credit card. expMonth: type: integer format: int32 description: Expiration month. expYear: type: integer format: int32 description: Expiration year. ListPaymentMethodsResponseStripePaymentMethod: type: object properties: id: type: string description: Payment method ID. card: $ref: "#/components/schemas/ListPaymentMethodsResponseCard" description: Card used by the payment method. usBankAccount: $ref: "#/components/schemas/ListPaymentMethodsResponseUsBankAccount" title: ACH Direct Debit ListPaymentMethodsResponseUsBankAccount: type: object properties: bankName: type: string last4: type: string description: Last four digits of the bank account number. ModelKind: type: string enum: - KIND_UNSPECIFIED - HF_BASE_MODEL - HF_PEFT_ADDON - HF_TEFT_ADDON - FLUMINA_BASE_MODEL - FLUMINA_ADDON - DRAFT_ADDON - LIVE_MERGE - CUSTOM_MODEL - EMBEDDING_MODEL - SNAPSHOT_MODEL default: KIND_UNSPECIFIED description: |- - HF_BASE_MODEL: An LLM base model. - HF_PEFT_ADDON: A parameter-efficent fine-tuned addon. - HF_TEFT_ADDON: A token-eficient fine-tuned addon. - FLUMINA_BASE_MODEL: A Flumina base model. - FLUMINA_ADDON: A Flumina addon. - DRAFT_ADDON: A draft model used for speculative decoding in a deployment. - LIVE_MERGE: A live-merge model. - CUSTOM_MODEL: A customized model - EMBEDDING_MODEL: An Embedding model. - SNAPSHOT_MODEL: A snapshot model. ModelSnapshotType: type: string enum: - FULL_SNAPSHOT - INCREMENTAL_SNAPSHOT default: FULL_SNAPSHOT PolicySettingsModelAccessRule: type: object properties: model: type: string permissions: $ref: "#/components/schemas/PolicySettingsModelPermissions" description: >- One per-model override. This is not a partial patch; it is a complete row replacing defaults for this model. required: - model - permissions PolicySettingsModelPermissions: type: object properties: allowServerless: type: boolean allowCoding: type: boolean allowDedicatedDeployments: type: boolean allowTraining: type: boolean description: >- Each boolean is required when ModelPermissions is sent. Using `optional` gives the field presence tracking, and `(buf.validate.field).required` makes server-side validation reject messages with any unset boolean. This forces callers (frontend, firectl, scripts) to always send all four permissions explicitly, instead of silently relying on bool's default of `false`. It also causes proto3 JSON to emit explicit `false` values on output, so a Get response always shows the full picture. required: - allowServerless - allowCoding - allowDedicatedDeployments - allowTraining PricingPlanAcceleratorHourBillingConfig: type: object properties: acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" title: Specific accelerator type this applies to acceleratorHourPrice: $ref: "#/components/schemas/typeMoney" title: Price per accelerator hour title: Configuration for accelerator-hour billing required: - acceleratorType - acceleratorHourPrice PricingPlanBillingType: type: string enum: - BILLING_TYPE_UNSPECIFIED - TOKEN_USAGE - ACCELERATOR_HOURS default: BILLING_TYPE_UNSPECIFIED PricingPlanTokenBillingConfig: type: object properties: baseModelName: type: string title: Model name this applies to inputTokenPricePerMillion: $ref: "#/components/schemas/typeMoney" title: Price per million input tokens outputTokenPricePerMillion: $ref: "#/components/schemas/typeMoney" title: Price per million output tokens uncachedInputTokenPricePerMillion: $ref: "#/components/schemas/typeMoney" title: Price per million uncached input tokens cachedInputTokenPricePerMillion: $ref: "#/components/schemas/typeMoney" title: Price per million cached input tokens title: Configuration for token-based billing required: - baseModelName - outputTokenPricePerMillion ReinforcementLearningLossConfigMethod: type: string enum: - METHOD_UNSPECIFIED - GRPO - DAPO - DPO - ORPO - GSPO_TOKEN default: METHOD_UNSPECIFIED title: |- - METHOD_UNSPECIFIED: Defaults to GRPO - GRPO: Group Relative Policy Optimization (default for preference jobs) - DAPO: Decoupled Alignment Preference Optimization - DPO: Direct Preference Optimization - ORPO: Odds Ratio Preference Optimization (reference-free) - GSPO_TOKEN: Group Sequence Policy Optimization (token-level) TrainingSessionReferenceState: type: string enum: - TRAINING_SESSION_REFERENCE_STATE_UNSPECIFIED - BASE - ADAPTER default: TRAINING_SESSION_REFERENCE_STATE_UNSPECIFIED description: |- ReferenceState indicates whether the session currently references the base model or a loaded LoRA adapter. Updated automatically when LoadTrainingSessionState is called. - BASE: Session is using the base model (no adapter loaded). - ADAPTER: Session has a LoRA adapter loaded. TrainingShapeTrainerMode: type: string enum: - TRAINER_MODE_UNSPECIFIED - POLICY_TRAINER - FORWARD_ONLY - LORA_TRAINER default: TRAINER_MODE_UNSPECIFIED description: Trainer execution mode used for validated launch-profile matching. UserUsageLimitIncreaseRequestRequestState: type: string enum: - REQUEST_STATE_UNSPECIFIED - PENDING - APPROVED - REJECTED default: REQUEST_STATE_UNSPECIFIED description: Lifecycle state of the request. ValidateAssertionsResponseValidateAssertionError: type: object properties: errorMessages: type: array items: type: string gatewayAcceleratorType: type: string enum: - ACCELERATOR_TYPE_UNSPECIFIED - NVIDIA_A100_80GB - NVIDIA_H100_80GB - AMD_MI300X_192GB - NVIDIA_A10G_24GB - NVIDIA_A100_40GB - NVIDIA_L4_24GB - NVIDIA_H200_141GB - NVIDIA_B200_180GB - AMD_MI325X_256GB - AMD_MI350X_288GB - NVIDIA_B300_288GB - NVIDIA_GB200 - NVIDIA_GB300 default: ACCELERATOR_TYPE_UNSPECIFIED gatewayAccount: type: object properties: name: type: string title: The resource name of the account. e.g. accounts/my-account readOnly: true displayName: type: string description: |- Human-readable display name of the account. e.g. "My Account" Must be fewer than 64 characters long. createTime: type: string format: date-time description: The creation time of the account. readOnly: true accountType: $ref: "#/components/schemas/AccountAccountType" description: The type of the account. email: type: string description: |- The primary email for the account. This is used for billing invoices and account notifications. state: $ref: "#/components/schemas/gatewayAccountState" description: The state of the account. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the account status. readOnly: true suspendState: $ref: "#/components/schemas/AccountSuspendState" readOnly: true updateTime: type: string format: date-time description: The update time for the account. readOnly: true notificationSettings: $ref: "#/components/schemas/gatewayNotificationSettings" description: Notification settings for this account. productivityOnly: type: boolean description: >- Marks this account as productivity/coding-only. Set once at account creation. required: - email gatewayAccountRateLimit: type: object properties: deployment: type: string title: >- Resource name of the deployment, e.g. accounts/my-account/deployments/my-deployment metric: $ref: "#/components/schemas/AccountRateLimitMetric" effectiveLimit: type: number format: double description: >- Enforced limit for this metric (per second; units depend on the metric). required: - deployment - metric - effectiveLimit gatewayAccountState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - UPDATING - DELETING default: STATE_UNSPECIFIED gatewayAccountUsage: type: object properties: serverlessCosts: type: array items: type: object $ref: "#/components/schemas/AccountUsageServerlessUsage" title: List of serverless cost data dedicatedCosts: type: array items: type: object $ref: "#/components/schemas/AccountUsageDedicatedDeploymentUsage" title: List of dedicated deployment cost data trainingCosts: type: array items: type: object $ref: "#/components/schemas/AccountUsageTrainingUsage" description: >- List of training costs (SFT / DPO token usage and RFT / service-mode accelerator-second usage). Empty when training usage was not requested or when the account has training usage hidden. title: Response with model costs by deployment type gatewayAccountUsageLimits: type: object properties: name: type: string title: |- Resource name of the account usage-limits singleton. Format: accounts/{account_id}/usageLimits readOnly: true defaultUserLimit: $ref: "#/components/schemas/typeMoney" description: >- The default per-user cap applied to every user without an override. Unset means no default cap (users are unlimited unless individually capped). enforcementMode: $ref: "#/components/schemas/gatewayAccountUsageLimitsEnforcementMode" description: Enforcement mode applied to the per-user caps in this account. used: $ref: "#/components/schemas/typeMoney" description: >- Account-wide current-period usage (the sum across users). Computed at read time; not stored. readOnly: true updateTime: type: string format: date-time description: >- The time the account usage limits (default cap / enforcement mode) were last updated. Unset until the account row is first written. readOnly: true description: >- AccountUsageLimits is the account-level usage-limits singleton: the default per-user cap and enforcement mode applied to every user in the account, plus the account-wide current-period usage rollup. There is exactly one per account. The default cap and enforcement mode are stored; the usage rollup is computed at read time. gatewayAccountUsageLimitsEnforcementMode: type: string enum: - ENFORCEMENT_MODE_UNSPECIFIED - DISABLED - WARN - BLOCK default: ENFORCEMENT_MODE_UNSPECIFIED description: |- EnforcementMode controls what happens when a user reaches their limit. - DISABLED: No tracking, no notifications, no enforcement. - WARN: Usage is tracked but never blocked. A safe stage before BLOCK. - BLOCK: Usage is tracked and the user is blocked at their limit until the period resets. gatewayApiKey: type: object properties: keyId: type: string description: >- Unique identifier (Key ID) for the API key, used primarily for deletion. readOnly: true displayName: type: string description: >- Display name for the API key, defaults to "default" if not specified. key: type: string description: >- The actual API key value, only available upon creation and not stored thereafter. readOnly: true createTime: type: string format: date-time description: Timestamp indicating when the API key was created. readOnly: true secure: type: boolean description: >- Indicates whether the plaintext value of the API key is unknown to Fireworks. If true, Fireworks does not know this API key's plaintext value. If false, Fireworks does know the plaintext value. readOnly: true email: type: string description: Email of the user who owns this API key. readOnly: true prefix: type: string title: The first few characters of the API key to visually identify it readOnly: true expireTime: type: string format: date-time description: >- Timestamp indicating when the API key will expire. If not set, the key never expires. annotations: type: object additionalProperties: type: string description: Optional key-value annotations for this API key. lastUsed: type: string format: date-time description: >- Timestamp indicating when the API key was last used. Not set if the key has never been used. readOnly: true isFirepass: type: boolean description: True when this key is a FirePass inference key. gatewayAssertion: type: object properties: assertionType: $ref: "#/components/schemas/AssertionAssertionType" llmAssertion: $ref: "#/components/schemas/gatewayLLMAssertion" codeAssertion: $ref: "#/components/schemas/gatewayCodeAssertion" metricName: type: string title: >- We are doing auto generated GORM with JSON serializer and oneof doesn't work so I am doing enums + just flat fields required: - assertionType gatewayAuditLogEntry: type: object properties: id: type: string description: Audit log entry id. method: type: string description: The gRPC method name. principal: type: string description: The email of the principal user who performed this action. payload: type: object description: The payload as JSON. status: $ref: "#/components/schemas/gatewayStatus" description: The response status. timestamp: type: string format: date-time description: The timestamp when the request was received. message: type: string title: Optional message describing the audit log entry resource: type: string title: The resource being operated on (e.g. accounts/123) isAdminAction: type: boolean title: >- Whether this action was taken by an admin and should not be shown to regular users userAgent: type: string description: >- The user agent that made the request (e.g. "firectl/1.2.3 grpc-go/1.73.0"). clientIp: type: string description: The client IP when known. apiKeyId: type: string description: >- Unkey API key ID when the action was authenticated with an API key (empty for bearer/OIDC). gatewayAutoTune: type: object properties: longPrompt: type: boolean description: If true, this deployment is optimized for long prompt lengths. gatewayAutoscalingPolicy: type: object properties: scaleUpWindow: type: string description: >- The duration the autoscaler will wait before scaling up a deployment after observing increased load. Default is 30s. Must be less than or equal to 1 hour. scaleDownWindow: type: string description: >- The duration the autoscaler will wait before scaling down a deployment after observing decreased load. Default is 10m. Must be less than or equal to 1 hour. scaleToZeroWindow: type: string description: >- The duration after which there are no requests that the deployment will be scaled down to zero replicas, if min_replica_count==0. Default is 1h. This must be at least 5 minutes. loadTargets: type: object additionalProperties: type: number format: float title: >- Map of load metric names to their target utilization factors. Currently only the "default" key is supported, which specifies the default target for all metrics. If not specified, the default target is 0.8 scalingSchedules: type: object additionalProperties: $ref: "#/components/schemas/gatewayScalingSchedule" description: >- Named scaling schedules that override min_replica_count on a time-based cron schedule. When multiple schedules are active simultaneously, the highest min_replica_count across all active schedules is used ("max wins"). When no schedule is active, the deployment's base min_replica_count applies. Maximum 5 schedules per deployment. gatewayAwsS3Config: type: object properties: credentialsSecret: type: string title: >- Reference to a Secret resource containing AWS access key credentials. Format: accounts/{account_id}/secrets/{secret_id} The secret value must be JSON: {"aws_access_key_id": "AKIA...", "aws_secret_access_key": "..."} iamRoleArn: type: string title: >- IAM role ARN to assume for accessing S3 datasets via GCP OIDC federation. Format: arn:aws:iam::account-id:role/role-name description: |- AwsS3Config is the configuration for AWS S3 dataset access which will be used by a training job. gatewayAwsS3ModelSource: type: object properties: s3Bucket: type: string description: The S3 bucket name. s3Path: type: string description: The S3 path prefix. roleArn: type: string title: AWS role ARN for authentication accessKeyId: type: string title: AWS access key ID for authentication accessSecret: type: string title: AWS access secret for authentication required: - s3Bucket gatewayAzureBlobModelSource: type: object properties: storageAccount: type: string description: The Azure storage account name. container: type: string description: The Azure blob container name. path: type: string description: The path prefix within the container. sasTokenSecret: type: string description: |- Reference to a Fireworks secret containing the Azure SAS token. Format: accounts/{account}/secrets/{secret_id} The secret's value should contain the SAS token. Create a secret first using the Secrets API, then reference it here. clientId: type: string description: |- Azure AD client ID for authentication. If set, tenant_id must also be set. tenantId: type: string description: |- Azure AD tenant ID for authentication. If set, client_id must also be set. required: - storageAccount - container gatewayAzureBlobStorageConfig: type: object properties: credentialsSecret: type: string description: >- Reference to a Secret resource containing Azure credentials. Format: accounts/{account_id}/secrets/{secret_id} The secret value must be JSON: {"connection_string": "..."} or {"sas_token": "..."} or {"account_key": "..."} Mutually exclusive with managed_identity_client_id. managedIdentityClientId: type: string description: >- Managed Identity Client ID for GCP-to-Azure Workload Identity Federation. Format: uuid Mutually exclusive with credentials_secret. tenantId: type: string title: |- Azure tenant ID for Workload Identity Federation. Format: uuid description: >- AzureBlobStorageConfig is the configuration for Azure Blob Storage dataset access which will be used by a training job. gatewayBalance: type: object properties: money: $ref: "#/components/schemas/typeMoney" description: >- Current remaining balance of the developer for a particular currency. gatewayBaseModelDetails: type: object properties: worldSize: type: integer format: int32 description: |- The default number of GPUs the model is served with. If not specified, the default is 1. checkpointFormat: $ref: "#/components/schemas/BaseModelDetailsCheckpointFormat" huggingfaceFiles: type: array items: type: string description: >- A list of Hugging Face files associated with this model. Specified if and only if the checkpoint_format is HUGGINGFACE. parameterCount: type: string format: int64 description: >- The number of model parameters. For serverless models, this determines the price per token. moe: type: boolean description: >- If true, this is a Mixture of Experts (MoE) model. For serverless models, this affects the price per token. tunable: type: boolean description: >- Deprecated: V1 training stack only. Use per-category tunable flags on Model instead. modelType: type: string description: The type of the model. supportsFireattention: type: boolean description: Whether this model supports fireattention. defaultPrecision: $ref: "#/components/schemas/DeploymentPrecision" description: Default precision of the model. readOnly: true supportsMtp: type: boolean description: If true, this model supports MTP. gatewayBaseTrainingConfig: type: object properties: outputModel: type: string description: >- The model ID to be assigned to the resulting fine-tuned model. If not specified, the job ID will be used. baseModel: type: string description: |- The name of the base model to be fine-tuned Only one of 'base_model' or 'warm_start_from' should be specified. warmStartFrom: type: string description: |- The PEFT addon model in Fireworks format to be fine-tuned from Only one of 'base_model' or 'warm_start_from' should be specified. jinjaTemplate: type: string description: >- Deprecated: literal Jinja templates are not supported by Training V2. Conversation rendering is selected from the base model's registered renderer configuration instead. learningRate: type: number format: float description: The learning rate used for training. maxContextLength: type: integer format: int32 description: The maximum context length to use with the model. loraRank: type: integer format: int32 description: The rank of the LoRA layers. epochs: type: integer format: int32 description: The number of epochs to train for. batchSize: type: integer format: int32 description: >- Deprecated: legacy V1 token budget. Training V2 batches by samples via batch_size_samples. gradientAccumulationSteps: type: integer format: int32 description: |- Deprecated: legacy V1 gradient accumulation. Training V2 batches by samples via batch_size_samples and rejects this field when set. learningRateWarmupSteps: type: integer format: int32 title: Number of steps for learning rate warm up batchSizeSamples: type: integer format: int32 description: The number of samples per gradient batch. optimizerWeightDecay: type: number format: float description: Weight decay (L2 regularization) for optimizer. trainerShardingScheme: $ref: "#/components/schemas/gatewayTrainerShardingScheme" description: Structured trainer sharding/parallelism configuration. loraAlpha: type: integer format: int32 description: |- LoRA alpha scaling factor. If not specified (or 0), trainer defaults are used. loraDropout: type: number format: float description: LoRA dropout probability. loraTargetModules: type: array items: type: string description: Optional LoRA target module names (e.g. q_proj, k_proj, v_proj). lrScheduler: $ref: "#/components/schemas/gatewayLearningRateScheduler" description: >- The learning-rate schedule (constant/linear/cosine + per-type knobs). When unset, the trainer uses the legacy constant schedule. title: |- BaseTrainingConfig contains common configuration fields shared across different training job types. gatewayBatchInferenceJob: type: object properties: name: type: string title: >- The resource name of the batch inference job. e.g. accounts/my-account/batchInferenceJobs/my-batch-inference-job readOnly: true displayName: type: string title: >- Human-readable display name of the batch inference job. e.g. "My Batch Inference Job" createTime: type: string format: date-time description: The creation time of the batch inference job. readOnly: true expireTime: type: string format: date-time description: >- The time when the batch inference job will expire (stop running); any completed requests will have been written to the output dataset by then. This is the job's effective execution deadline, derived by the server as create_time + the bounded run window (see max_job_duration). It is exposed so customers can read back the concrete deadline without recomputing it client-side. OUTPUT_ONLY: it is always computed server-side and any client-supplied value is ignored (previously this was a SUPERUSER_ONLY input that overlapped with max_job_duration; the two are now unified as one public input (duration) + one public derived deadline (this timestamp)). readOnly: true createdBy: type: string description: >- The email address of the user who initiated this batch inference job. readOnly: true state: $ref: "#/components/schemas/gatewayJobState" description: JobState represents the state an asynchronous job can be in. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true model: type: string description: >- The name of the model to use for inference. This is required, except when continued_from_job_name is specified. inputDatasetId: type: string description: >- The name of the dataset used for inference. This is required, except when continued_from_job_name is specified. outputDatasetId: type: string description: >- The name of the dataset used for storing the results. This will also contain the error file. systemPrompt: type: string description: >- Optional job-level system prompt. When set, it is injected as a leading system message into every input row that does NOT already begin with a system message (a row's own leading system message takes precedence). This lets callers avoid repeating a large, static system prompt on every row of the input dataset, shrinking the upload. Because the injected prefix is byte-identical across rows, prompt caching still applies. inferenceParameters: $ref: "#/components/schemas/gatewayBatchInferenceJobInferenceParameters" description: Parameters controlling the inference process. updateTime: type: string format: date-time description: The update time for the batch inference job. readOnly: true precision: $ref: "#/components/schemas/DeploymentPrecision" description: >- The precision with which the model should be served. If PRECISION_UNSPECIFIED, a default will be chosen based on the model. jobProgress: $ref: "#/components/schemas/gatewayJobProgress" description: Job progress. readOnly: true continuedFromJobName: type: string description: >- The resource name of the batch inference job that this job continues from. Used for lineage tracking to understand job continuation chains. placement: $ref: "#/components/schemas/gatewayPlacement" description: >- The desired geographic region where the batch inference job runs. Set `multi_region` to limit the job to a region group (US, EUROPE, APAC, or GLOBAL). If unspecified, the job runs in any supported region. maxJobDuration: type: string description: >- The customer-requested wall-clock run window for the job: how long it may run before it is expired. This is the single public input that controls the job's lifetime. The server bounds it to [12h, 72h]; if unset it defaults to 24h. The resulting concrete deadline is surfaced as expire_time (= create_time + the bounded window) and the job is expired once that deadline passes. A duration (relative) is used rather than an absolute timestamp because the client does not know create_time at submit time. Customer-visible input. lifecycle: $ref: "#/components/schemas/BatchInferenceJobLifecycleTimestamps" description: >- Lifecycle milestone timestamps (validated / run-start / end) for the job. readOnly: true waitingOnCapacity: type: boolean description: >- True only while a job that has ALREADY started running is briefly re-acquiring capacity after a mid-run preemption/stockout (i.e. it regressed from RUNNING back to an internal PENDING/CREATING phase and is waiting to resume). This is intentionally a transient sub-status annotation on a job whose customer-facing `state` stays RUNNING — NOT a distinct `state` value: the job is still running (progress is saved, it auto-resumes) so introducing a new terminal-or-not enum value would force every state consumer (SDK/CLI/internal maps/billing) to special-case "still running." It drives the customer-facing "Briefly paused — waiting on capacity" card. It must NOT be set during first-time provisioning before the job has ever run, and is cleared once the job returns to RUNNING or reaches a terminal state. So "state=RUNNING, waiting_on_capacity=true" means: running, momentarily paused while it re-acquires capacity. readOnly: true gatewayBatchInferenceJobInferenceParameters: type: object properties: maxTokens: type: integer format: int32 description: Maximum number of tokens to generate per response. temperature: type: number format: float description: Sampling temperature, typically between 0 and 2. topP: type: number format: float description: Top-p sampling parameter, typically between 0 and 1. "n": type: integer format: int32 description: Number of response candidates to generate per input. extraBody: type: string description: |- Additional parameters for the inference request as a JSON string. For example: "{\"stop\": [\"\\n\"]}". topK: type: integer format: int32 description: >- Top-k sampling parameter, limits the token selection to the top k tokens. title: BIJ inference parameters gatewayBillcomInvoice: type: object properties: id: type: string description: The Bill.com invoice ID (e.g. "00e02ZLYJNYUXC216neu"). invoiceNumber: type: string title: The Bill.com invoice number (e.g. "167") archived: type: boolean description: Whether this invoice is archived. invoiceDate: $ref: "#/components/schemas/typeDate" description: The date when the invoice was issued (calendar date, no timezone). dueDate: $ref: "#/components/schemas/typeDate" description: The date when payment is due (calendar date, no timezone). customerId: type: string description: The Bill.com customer ID. totalAmount: $ref: "#/components/schemas/typeMoney" description: The total invoice amount. dueAmount: $ref: "#/components/schemas/typeMoney" description: The amount still due (after payments/credits). scheduledAmount: $ref: "#/components/schemas/typeMoney" description: The amount scheduled for payment. creditAmount: $ref: "#/components/schemas/typeMoney" description: The amount of credit amount applied. status: type: string description: The invoice status (DRAFT, OPEN, PAID, VOID, etc.). salesTaxTotal: $ref: "#/components/schemas/typeMoney" description: The total amount of sales tax. paymentLink: type: string description: A shareable payment link for viewing and paying this invoice. title: Bill.com invoice (actual invoices for enterprise contracts) gatewayBillingRequestCostRecord: type: object properties: requestId: type: string title: Request ID costNanoUsd: type: number format: double title: Cost title: Individual billing request cost Record gatewayBillingSettings: type: object properties: name: type: string title: Resource name, e.g. accounts/my-account/billingSettings readOnly: true updateTime: type: string format: date-time description: The time the billing settings were last updated. readOnly: true enableCreditAutoReload: type: boolean description: If true, the account will auto-reload its credit balance. creditAutoReloadThreshold: $ref: "#/components/schemas/typeMoney" description: |- The threshold at which to auto-reload the account's credit balance. Must be less than $100,000. creditAutoReloadAmount: $ref: "#/components/schemas/typeMoney" description: |- The amount to auto-reload the account's credit balance. Must be less than $100,000. gatewayCheckpointAvailability: type: object properties: availabilityClass: $ref: "#/components/schemas/CheckpointAvailabilityAvailabilityClass" readOnly: true region: $ref: "#/components/schemas/gatewayRegion" description: >- The region containing this checkpoint copy. Managed durable availability may leave this unset when the backing storage is not tied to one runtime region. readOnly: true createTime: type: string format: date-time readOnly: true updateTime: type: string format: date-time readOnly: true gatewayCheckpointType: type: string enum: - CHECKPOINT_TYPE_UNSPECIFIED - CHECKPOINT_TYPE_INFERENCE_BASE - CHECKPOINT_TYPE_INFERENCE_ARC_V2 - CHECKPOINT_TYPE_INFERENCE_LORA - CHECKPOINT_TYPE_TRAINING - CHECKPOINT_TYPE_TRAINING_LORA default: CHECKPOINT_TYPE_UNSPECIFIED description: >- - CHECKPOINT_TYPE_INFERENCE_BASE: Inference-ready checkpoint (full-parameter, HuggingFace format). Promotable to a model for serving. - CHECKPOINT_TYPE_INFERENCE_ARC_V2: Inference-ready checkpoint (XOR delta compressed, ARC v2 format). Not directly promotable. - CHECKPOINT_TYPE_INFERENCE_LORA: Inference-ready LoRA adapter checkpoint (HuggingFace PEFT format). Promotable to an addon model for serving. - CHECKPOINT_TYPE_TRAINING: Training checkpoint (full-parameter) with optimizer state. Not promotable to a model; used for training resume. - CHECKPOINT_TYPE_TRAINING_LORA: Training checkpoint from a LoRA run (adapter weights + optimizer state). Not promotable; used for training resume. gatewayCluster: type: object properties: name: type: string title: >- The resource name of the cluster. e.g. accounts/my-account/clusters/my-cluster readOnly: true displayName: type: string description: |- Human-readable display name of the cluster. e.g. "My Cluster" Must be fewer than 64 characters long. createTime: type: string format: date-time description: The creation time of the cluster. readOnly: true eksCluster: $ref: "#/components/schemas/gatewayEksCluster" fakeCluster: $ref: "#/components/schemas/gatewayFakeCluster" state: $ref: "#/components/schemas/gatewayClusterState" description: The current state of the cluster. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Detailed information about the current status of the cluster. readOnly: true updateTime: type: string format: date-time description: The update time for the cluster. readOnly: true gatewayClusterConnectionInfo: type: object properties: endpoint: type: string description: The cluster's Kubernetes API server endpoint. caData: type: string description: Base64-encoded cluster's CA certificate. gatewayClusterState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - DELETING - FAILED default: STATE_UNSPECIFIED description: |- - CREATING: The cluster is still being created. - READY: The cluster is ready to be used. - DELETING: The cluster is being deleted. - FAILED: Cluster is not operational. Consult 'status' for detailed messaging. Cluster needs to be deleted and re-created. gatewayCode: type: string enum: - OK - CANCELLED - UNKNOWN - INVALID_ARGUMENT - DEADLINE_EXCEEDED - NOT_FOUND - ALREADY_EXISTS - PERMISSION_DENIED - UNAUTHENTICATED - RESOURCE_EXHAUSTED - FAILED_PRECONDITION - ABORTED - OUT_OF_RANGE - UNIMPLEMENTED - INTERNAL - UNAVAILABLE - DATA_LOSS default: OK description: |- - OK: Not an error; returned on success. HTTP Mapping: 200 OK - CANCELLED: The operation was cancelled, typically by the caller. HTTP Mapping: 499 Client Closed Request - UNKNOWN: Unknown error. For example, this error may be returned when a `Status` value received from another address space belongs to an error space that is not known in this address space. Also errors raised by APIs that do not return enough error information may be converted to this error. HTTP Mapping: 500 Internal Server Error - INVALID_ARGUMENT: The client specified an invalid argument. Note that this differs from `FAILED_PRECONDITION`. `INVALID_ARGUMENT` indicates arguments that are problematic regardless of the state of the system (e.g., a malformed file name). HTTP Mapping: 400 Bad Request - DEADLINE_EXCEEDED: The deadline expired before the operation could complete. For operations that change the state of the system, this error may be returned even if the operation has completed successfully. For example, a successful response from a server could have been delayed long enough for the deadline to expire. HTTP Mapping: 504 Gateway Timeout - NOT_FOUND: Some requested entity (e.g., file or directory) was not found. Note to server developers: if a request is denied for an entire class of users, such as gradual feature rollout or undocumented allowlist, `NOT_FOUND` may be used. If a request is denied for some users within a class of users, such as user-based access control, `PERMISSION_DENIED` must be used. HTTP Mapping: 404 Not Found - ALREADY_EXISTS: The entity that a client attempted to create (e.g., file or directory) already exists. HTTP Mapping: 409 Conflict - PERMISSION_DENIED: The caller does not have permission to execute the specified operation. `PERMISSION_DENIED` must not be used for rejections caused by exhausting some resource (use `RESOURCE_EXHAUSTED` instead for those errors). `PERMISSION_DENIED` must not be used if the caller can not be identified (use `UNAUTHENTICATED` instead for those errors). This error code does not imply the request is valid or the requested entity exists or satisfies other pre-conditions. HTTP Mapping: 403 Forbidden - UNAUTHENTICATED: The request does not have valid authentication credentials for the operation. HTTP Mapping: 401 Unauthorized - RESOURCE_EXHAUSTED: Some resource has been exhausted, perhaps a per-user quota, or perhaps the entire file system is out of space. HTTP Mapping: 429 Too Many Requests - FAILED_PRECONDITION: The operation was rejected because the system is not in a state required for the operation's execution. For example, the directory to be deleted is non-empty, an rmdir operation is applied to a non-directory, etc. Service implementors can use the following guidelines to decide between `FAILED_PRECONDITION`, `ABORTED`, and `UNAVAILABLE`: (a) Use `UNAVAILABLE` if the client can retry just the failing call. (b) Use `ABORTED` if the client should retry at a higher level. For example, when a client-specified test-and-set fails, indicating the client should restart a read-modify-write sequence. (c) Use `FAILED_PRECONDITION` if the client should not retry until the system state has been explicitly fixed. For example, if an "rmdir" fails because the directory is non-empty, `FAILED_PRECONDITION` should be returned since the client should not retry unless the files are deleted from the directory. HTTP Mapping: 400 Bad Request - ABORTED: The operation was aborted, typically due to a concurrency issue such as a sequencer check failure or transaction abort. See the guidelines above for deciding between `FAILED_PRECONDITION`, `ABORTED`, and `UNAVAILABLE`. HTTP Mapping: 409 Conflict - OUT_OF_RANGE: The operation was attempted past the valid range. E.g., seeking or reading past end-of-file. Unlike `INVALID_ARGUMENT`, this error indicates a problem that may be fixed if the system state changes. For example, a 32-bit file system will generate `INVALID_ARGUMENT` if asked to read at an offset that is not in the range [0,2^32-1], but it will generate `OUT_OF_RANGE` if asked to read from an offset past the current file size. There is a fair bit of overlap between `FAILED_PRECONDITION` and `OUT_OF_RANGE`. We recommend using `OUT_OF_RANGE` (the more specific error) when it applies so that callers who are iterating through a space can easily look for an `OUT_OF_RANGE` error to detect when they are done. HTTP Mapping: 400 Bad Request - UNIMPLEMENTED: The operation is not implemented or is not supported/enabled in this service. HTTP Mapping: 501 Not Implemented - INTERNAL: Internal errors. This means that some invariants expected by the underlying system have been broken. This error code is reserved for serious errors. HTTP Mapping: 500 Internal Server Error - UNAVAILABLE: The service is currently unavailable. This is most likely a transient condition, which can be corrected by retrying with a backoff. Note that it is not always safe to retry non-idempotent operations. See the guidelines above for deciding between `FAILED_PRECONDITION`, `ABORTED`, and `UNAVAILABLE`. HTTP Mapping: 503 Service Unavailable - DATA_LOSS: Unrecoverable data loss or corruption. HTTP Mapping: 500 Internal Server Error title: >- Mimics [https://github.com/googleapis/googleapis/blob/master/google/rpc/code.proto] gatewayCodeAssertion: type: object properties: language: type: string title: Language of the code (python/javascript) code: type: string title: The code to execute expectedOutput: type: string title: Optional expected output options: $ref: "#/components/schemas/CodeAssertionExecutionOptions" required: - language - code gatewayCodeSnippets: type: object properties: language: type: string fileContents: type: object additionalProperties: type: string title: File name to code snippet, default is main.py entryFile: type: string entryFunc: type: string gatewayConstantLRSchedule: type: object description: >- Constant learning rate held flat after warmup (legacy default). No decay knobs. gatewayConversationConfig: type: object properties: style: type: string description: The chat template to use. system: type: string description: The system prompt (if the chat style supports it). template: type: string description: The Jinja template (if style is "jinja"). required: - style gatewayCorrelationMatrix: type: object properties: labels: type: array items: type: string rows: type: array items: type: object $ref: "#/components/schemas/gatewayCorrelationRow" description: |- CorrelationMatrix is a square matrix; rows[i].values[j] is the Pearson correlation between labels[i] and labels[j]. gatewayCorrelationRow: type: object properties: values: type: array items: type: number format: double gatewayCosineLRSchedule: type: object properties: minLrRatio: type: number format: float description: >- Floor learning rate as a fraction of the peak learning rate (0.0 = decay to zero, 0.1 = decay to 10% of the peak learning rate). decayRatio: type: number format: float description: >- Fraction of total training steps over which to decay. 0 (unset) decays over the full run. description: >- Cosine annealing from the peak learning rate toward min_lr_ratio after warmup. gatewayCreateDeploymentRequest: type: object properties: parent: type: string description: The resource name of the parent account. deployment: $ref: "#/components/schemas/gatewayDeployment" description: The properties of the deployment being created. disableAutoDeploy: type: boolean description: >- By default, a deployment created with a currently undeployed base model will be deployed to this deployment. If true, this auto-deploy function is disabled. disableSpeculativeDecoding: type: boolean description: >- By default, a deployment will use the speculative decoding settings from the base model. If true, this will disable speculative decoding. deploymentId: type: string description: >- The ID of the deployment. If not specified, a random ID will be generated. validateOnly: type: boolean description: >- If true, this will not create the deployment, but will return the deployment that would be created. skipShapeValidation: type: boolean description: >- By default, a deployment will ensure the deployment shape provided is validated. If true, we will not require the deployment shape to be validated. skipImageTagValidation: type: boolean description: >- If true, skip the image tag policy validation that blocks certain image tags. This allows creating deployments with image tags that would otherwise be blocked. required: - parent - deployment gatewayCreditRedemption: type: object properties: name: type: string description: The resource name of the credit redemption. readOnly: true creditCode: type: string title: The user-facing code of the credit code being redeemed createTime: type: string format: date-time description: >- The creation time of the credit redemption (i.e. when the credit code was redeemed). readOnly: true title: >- CreditRedemption represents a record of a code redemption. Stored as (code, account) pairs required: - creditCode gatewayCriterion: type: object properties: type: $ref: "#/components/schemas/gatewayCriterionType" name: type: string description: type: string codeSnippets: $ref: "#/components/schemas/gatewayCodeSnippets" title: Criteria for code snippet gatewayCriterionType: type: string enum: - TYPE_UNSPECIFIED - CODE_SNIPPETS default: TYPE_UNSPECIFIED title: "- CODE_SNIPPETS: Code snippets for Sandbox based evaluation" gatewayDataset: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true state: $ref: "#/components/schemas/gatewayDatasetState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true exampleCount: type: string format: int64 userUploaded: $ref: "#/components/schemas/gatewayUserUploaded" evaluationResult: $ref: "#/components/schemas/gatewayEvaluationResult" transformed: $ref: "#/components/schemas/gatewayTransformed" splitted: $ref: "#/components/schemas/gatewaySplitted" evalProtocol: $ref: "#/components/schemas/gatewayEvalProtocol" externalUrl: type: string title: The external URI of the dataset. e.g. gs://foo/bar/baz.jsonl format: $ref: "#/components/schemas/DatasetFormat" createdBy: type: string description: The email address of the user who initiated this fine-tuning job. readOnly: true updateTime: type: string format: date-time description: The update time for the dataset. readOnly: true sourceJobName: type: string description: >- The resource name of the job that created this dataset (e.g., batch inference job). Used for lineage tracking to understand dataset provenance. estimatedTokenCount: type: string format: int64 description: The estimated number of tokens in the dataset. readOnly: true averageTurnCount: type: number format: float description: >- An estimate of the average number of turns per sample in the dataset. readOnly: true encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: >- CMEK encryption state (authoritative, immutable, stamped at creation). readOnly: true gatewayDatasetState: type: string enum: - STATE_UNSPECIFIED - UPLOADING - READY default: STATE_UNSPECIFIED gatewayDatasetValidationFormat: type: string enum: - DATASET_VALIDATION_FORMAT_UNSPECIFIED - DATASET_VALIDATION_FORMAT_CHAT - DATASET_VALIDATION_FORMAT_CHAT_RELAXED - DATASET_VALIDATION_FORMAT_RLOR - DATASET_VALIDATION_FORMAT_BATCH_INFERENCE_STRICT - DATASET_VALIDATION_FORMAT_BATCH_INFERENCE_ALLOW_TRAILING_ASSISTANT - DATASET_VALIDATION_FORMAT_BATCH_INFERENCE - DATASET_VALIDATION_FORMAT_DPO default: DATASET_VALIDATION_FORMAT_UNSPECIFIED description: >- - DATASET_VALIDATION_FORMAT_UNSPECIFIED: Unspecified format. - DATASET_VALIDATION_FORMAT_CHAT: Chat dataset format. One or more .jsonl files, in which each line is a JSON object with a "messages" list, containing objects with a "role" field and a "content" field. {"messages": [{"role": "user", "content": "Hello, world!"}, {"role": "assistant", "content": "Hello, user!"}]} - DATASET_VALIDATION_FORMAT_CHAT_RELAXED: Chat dataset relaxed format. Same as DATASET_VALIDATION_FORMAT_CHAT, but optionally allows the assistant message to be missing, and allows extra fields. - DATASET_VALIDATION_FORMAT_RLOR: RLOR dataset format. One or more .jsonl files, in which each line is a JSON object with a "samples" array, containing objects with "messages" list and optional metrics keys with values ranging from 0.0 to 1.0. {"samples": [{"messages": [{"role": "user", "content": "..."}, {"role": "assistant", "content": "..."}], "metric_key": 1.0}, ...]} - DATASET_VALIDATION_FORMAT_BATCH_INFERENCE_STRICT: Batch inference strict format. Same as chat except no trailing assistant messages allowed. - DATASET_VALIDATION_FORMAT_BATCH_INFERENCE_ALLOW_TRAILING_ASSISTANT: Batch inference format with trailing assistant messages allowed. Same as chat except trailing assistant messages are optional. - DATASET_VALIDATION_FORMAT_BATCH_INFERENCE: Batch inference format (new). It will allow chat format and also OpenAI format. - DATASET_VALIDATION_FORMAT_DPO: DPO format (new). It will allow DPO format. gatewayDatasetValidationJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true datasetName: type: string description: The name of the dataset to validate. format: $ref: "#/components/schemas/gatewayDatasetValidationFormat" description: >- The format of the dataset. See DatasetValidationFormat for more details. result: $ref: "#/components/schemas/gatewayDatasetValidationJobResult" readOnly: true validationError: type: string readOnly: true datasetNames: type: array items: type: string description: The name of the dataset to validate. updateTime: type: string format: date-time description: The update time for the dataset validation job. readOnly: true rewards: type: array items: type: string description: A list of reward metrics to validate. baseModel: type: string description: >- Resource name of the base model associated with the parent job. This is used to determine the tokenizer. skipChatTemplateValidation: type: boolean description: If set, chat template validation is disabled. required: - datasetName - format gatewayDatasetValidationJobResult: type: string enum: - DATASET_VALIDATION_JOB_RESULT_UNSPECIFIED - DATASET_VALIDATION_JOB_RESULT_SUCCESS - DATASET_VALIDATION_JOB_RESULT_FAILURE default: DATASET_VALIDATION_JOB_RESULT_UNSPECIFIED gatewayDebugReinforcementFineTuningJobResponse: type: object properties: name: type: string description: The resource name of the reinforcement fine-tuning job. failedJobName: type: string readOnly: true required: - name gatewayDeployedModel: type: object properties: name: type: string title: >- The resource name. e.g. accounts/my-account/deployedModels/my-deployed-model readOnly: true displayName: type: string description: type: string description: Description of the resource. createTime: type: string format: date-time description: The creation time of the resource. readOnly: true model: type: string title: |- The resource name of the model to be deployed. e.g. accounts/my-account/models/my-model deployment: type: string description: The resource name of the base deployment the model is deployed to. default: type: boolean description: >- If true, this is the default target when querying this model without the `#` suffix. The first deployment a model is deployed to will have this field set to true. state: $ref: "#/components/schemas/gatewayDeployedModelState" description: The state of the deployed model. readOnly: true serverless: type: boolean title: True if the underlying deployment is managed by Fireworks status: $ref: "#/components/schemas/gatewayStatus" description: Contains model deploy/undeploy details. readOnly: true public: type: boolean description: If true, the deployed model will be publicly reachable. updateTime: type: string format: date-time description: The update time for the deployed model. readOnly: true gatewayDeployedModelRef: type: object properties: name: type: string title: >- The resource name. e.g. accounts/my-account/deployedModels/my-deployed-model readOnly: true deployment: type: string description: The resource name of the base deployment the model is deployed to. readOnly: true state: $ref: "#/components/schemas/gatewayDeployedModelState" description: The state of the deployed model. readOnly: true default: type: boolean description: >- If true, this is the default target when querying this model without the `#` suffix. The first deployment a model is deployed to will have this field set to true automatically. readOnly: true public: type: boolean description: If true, the deployed model will be publicly reachable. readOnly: true gatewayDeployedModelState: type: string enum: - STATE_UNSPECIFIED - UNDEPLOYING - DEPLOYING - DEPLOYED - UPDATING default: STATE_UNSPECIFIED description: |- - UNDEPLOYING: The model is being undeployed. - DEPLOYING: The model is being deployed. - DEPLOYED: The model is deployed and ready for inference. - UPDATING: there are updates happening with the deployed model gatewayDeployment: type: object properties: name: type: string title: >- The resource name of the deployment. e.g. accounts/my-account/deployments/my-deployment readOnly: true displayName: type: string description: |- Human-readable display name of the deployment. e.g. "My Deployment" Must be fewer than 64 characters long. description: type: string description: Description of the deployment. createTime: type: string format: date-time description: The creation time of the deployment. readOnly: true expireTime: type: string format: date-time description: >- Deprecated: This field is deprecated and no longer causes auto-deletion. The time at which this deployment will automatically be deleted. purgeTime: type: string format: date-time description: The time at which the resource will be hard deleted. readOnly: true deleteTime: type: string format: date-time description: The time at which the resource will be soft deleted. readOnly: true state: $ref: "#/components/schemas/gatewayDeploymentState" description: The state of the deployment. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Detailed status information regarding the most recent operation. readOnly: true annotations: type: object additionalProperties: type: string description: >- Annotations to identify deployment properties. Key/value pairs may be used by external tools or other services. The "image-tag-reason" key is redacted from API responses for non-superuser principals. minReplicaCount: type: integer format: int32 description: |- The minimum number of replicas. If not specified, the default is 0. maxReplicaCount: type: integer format: int32 description: |- The maximum number of replicas. If not specified, the default is max(min_replica_count, 1). May be set to 0 to downscale the deployment to 0. maxWithRevocableReplicaCount: type: integer format: int32 description: >- max_with_revocable_replica_count is max replica count including revocable capacity. The max revocable capacity will be max_with_revocable_replica_count - max_replica_count. desiredReplicaCount: type: integer format: int32 description: >- The desired number of replicas for this deployment. This represents the target replica count that the system is trying to achieve. readOnly: true replicaCount: type: integer format: int32 readOnly: true autoscalingPolicy: $ref: "#/components/schemas/gatewayAutoscalingPolicy" baseModel: type: string title: The base model name. e.g. accounts/fireworks/models/falcon-7b acceleratorCount: type: integer format: int32 description: >- The number of accelerators used per replica. If not specified, the default is the estimated minimum required by the base model. acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: The type of accelerator to use. precision: $ref: "#/components/schemas/DeploymentPrecision" description: The precision with which the model should be served. maxConcurrencyPerReplica: type: integer format: int32 description: >- The maximum number of concurrent (in-flight) requests a single replica will accept before shedding load. Requests that arrive while a replica is already at this limit are rejected early with HTTP 429 instead of queueing — a per-replica admission gate for controlling tail latency. When unset (0), the platform default is used. cluster: type: string description: If set, this deployment is deployed to a cloud-premise cluster. readOnly: true enableAddons: type: boolean description: If true, PEFT addons are enabled for this deployment. draftTokenCount: type: integer format: int32 description: >- The number of candidate tokens to generate per step for speculative decoding. Default is the base model's draft_token_count. Set CreateDeploymentRequest.disable_speculative_decoding to false to disable this behavior. draftModel: type: string description: >- The draft model name for speculative decoding. e.g. accounts/fireworks/models/my-draft-model If empty, speculative decoding using a draft model is disabled. Default is the base model's default_draft_model. Set CreateDeploymentRequest.disable_speculative_decoding to false to disable this behavior. ngramSpeculationLength: type: integer format: int32 description: >- The length of previous input sequence to be considered for N-gram speculation. enableSessionAffinity: type: boolean description: |- Whether to apply sticky routing based on `user` field. Serverless will be set to true when creating deployment. directRouteApiKeys: type: array items: type: string description: >- The set of API keys used to access the direct route deployment. If direct routing is not enabled, this field is unused. numPeftDeviceCached: type: integer format: int32 title: How many peft adapters to keep on gpu side for caching directRouteType: $ref: "#/components/schemas/gatewayDirectRouteType" description: >- If set, this deployment will expose an endpoint that bypasses the Fireworks API gateway. directRouteHandle: type: string description: >- The handle for calling a direct route. The meaning of the handle depends on the direct route type of the deployment: INTERNET -> The host name for accessing the deployment GCP_PRIVATE_SERVICE_CONNECT -> The service attachment name used to create the PSC endpoint. AWS_PRIVATELINK -> The service name used to create the VPC endpoint. readOnly: true deploymentTemplate: type: string description: |- The name of the deployment template to use for this deployment. Only available to enterprise accounts. autoTune: $ref: "#/components/schemas/gatewayAutoTune" description: The performance profile to use for this deployment. placement: $ref: "#/components/schemas/gatewayPlacement" description: |- The desired geographic region where the deployment must be placed. If unspecified, the default is the GLOBAL multi-region. region: $ref: "#/components/schemas/gatewayRegion" description: >- The geographic region where the deployment is presently located. This region may change over time, but within the `placement` constraint. readOnly: true maxContextLength: type: integer format: int32 description: >- The maximum context length supported by the model (context window). If set to 0 or not specified, the model's default maximum context length will be used. updateTime: type: string format: date-time description: The update time for the deployment. readOnly: true disableDeploymentSizeValidation: type: boolean description: Whether the deployment size validation is disabled. enableHotLoad: type: boolean description: Whether to use hot load for this deployment. hotLoadBucketType: $ref: "#/components/schemas/DeploymentHotLoadBucketType" title: >- hot load bucket name, indicate what type of storage to use for hot load enableHotReloadLatestAddon: type: boolean description: >- Allows up to 1 addon at a time to be loaded, and will merge it into the base model. deploymentShape: type: string description: >- The name of the deployment shape that this deployment is using. On the server side, this will be replaced with the deployment shape version name. activeModelVersion: type: string description: >- The model version that is currently active and applied to running replicas of a deployment. targetModelVersion: type: string description: >- The target model version that is being rolled out to the deployment. In a ready steady state, the target model version is the same as the active model version. replicaStats: $ref: "#/components/schemas/gatewayReplicaStats" description: >- Per-replica deployment status counters. Provides visibility into the deployment process by tracking replicas in different stages of the deployment lifecycle. readOnly: true hotLoadBucketUrl: type: string title: >- For hot load bucket location e.g for s3: s3://mybucket/; for GCS: gs://mybucket/, no trailing slash pricingPlanId: type: string description: |- Optional pricing plan ID for custom billing configuration. If set, this deployment will use the pricing plan's billing rules instead of default billing behavior. hotLoadTrainerJob: type: string title: >- Trainer job whose hot_load_bucket_url this deployment should use. At creation, the trainer's bucket URL is copied into this deployment's hot_load_bucket_url. The deployment continues working after the trainer is deleted (snapshot-at-creation semantics). Format: accounts/{account}/rlorTrainerJobs/{job} hotLoadTransitionType: $ref: "#/components/schemas/DeploymentHotLoadTransitionType" description: >- Controls hot-load reload semantics for this deployment. ASYNC (default when enable_hot_load is set) pauses the generator/prefiller mid-flight and skips draining in-flight requests during hot load. SYNC drains in-flight requests before applying the new model. preemptible: type: boolean description: When true, this deployment runs as preemptible. required: - baseModel gatewayDeploymentAcceleratorConfig: type: object properties: acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" title: "Key: The accelerator type for this config" precision: $ref: "#/components/schemas/DeploymentPrecision" description: "Key: Supported precisions." minAcceleratorCount: type: integer format: int32 description: "Value: Minimum number of accelerators required." regions: type: array items: $ref: "#/components/schemas/gatewayRegion" description: >- Value: Regions where the account has quota for this accelerator type. multiRegions: type: array items: $ref: "#/components/schemas/gatewayMultiRegion" description: >- Value: Multi-regions where the account has quota for this accelerator type. gatewayDeploymentPrerequisites: type: object properties: acceleratorConfigs: type: array items: type: object $ref: "#/components/schemas/gatewayDeploymentAcceleratorConfig" gatewayDeploymentShape: type: object properties: name: type: string title: >- The resource name of the deployment shape. e.g. accounts/my-account/deploymentShapes/my-deployment-shape readOnly: true displayName: type: string description: >- Human-readable display name of the deployment shape. e.g. "My Deployment Shape" Must be fewer than 64 characters long. description: type: string description: >- The description of the deployment shape. Must be fewer than 1000 characters long. createTime: type: string format: date-time description: The creation time of the deployment shape. readOnly: true updateTime: type: string format: date-time description: The update time for the deployment shape. readOnly: true baseModel: type: string description: >- Mutable, but only via UpdateDeploymentShape and only to a compatible model: one with the same model_type, parameter_count, and embedding status. The derived model_type / parameter_count fields (OUTPUT_ONLY, IMMUTABLE) are therefore preserved across a swap. Any incompatible change is rejected. title: The base model name. e.g. accounts/fireworks/models/falcon-7b modelType: type: string description: The model type of the base model. readOnly: true parameterCount: type: string format: int64 description: The parameter count of the base model . readOnly: true acceleratorCount: type: integer format: int32 description: >- The number of accelerators used per replica. If not specified, the default is the estimated minimum required by the base model. acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: |- The type of accelerator to use. If not specified, the default is NVIDIA_A100_80GB. precision: $ref: "#/components/schemas/DeploymentPrecision" description: The precision with which the model should be served. disableDeploymentSizeValidation: type: boolean description: If true, the deployment size validation is disabled. enableAddons: type: boolean description: >- If true, LORA addons are enabled for deployments created from this shape. Deprecated: set enable_addons on the deployment instead. draftTokenCount: type: integer format: int32 description: |- The number of candidate tokens to generate per step for speculative decoding. Default is the base model's draft_token_count. draftModel: type: string description: >- The draft model name for speculative decoding. e.g. accounts/fireworks/models/my-draft-model If empty, speculative decoding using a draft model is disabled. Default is the base model's default_draft_model. Deprecated: set default_draft_model on the base model instead. ngramSpeculationLength: type: integer format: int32 description: >- The length of previous input sequence to be considered for N-gram speculation. disableSpeculativeDecoding: type: boolean description: >- DEPRECATED: This field is a no-op. Speculative decoding is configured on deployment. enableSessionAffinity: type: boolean description: Whether to apply sticky routing based on `user` field. numLoraDeviceCached: type: integer format: int32 title: How many LORA adapters to keep on GPU side for caching maxContextLength: type: integer format: int32 description: >- The maximum context length supported by the model (context window). If set to 0 or not specified, the model's default maximum context length will be used. presetType: $ref: "#/components/schemas/DeploymentShapePresetType" description: Type of deployment shape for different deployment configurations. title: >- A deployment shape is a set of parameters that define the shape of a deployment. Deployments are created from a deployment shape. required: - baseModel gatewayDeploymentShapeVersion: type: object properties: name: type: string title: >- The resource name of the deployment shape version. e.g. accounts/my-account/deploymentShapes/my-deployment-shape/versions/{version_id} readOnly: true createTime: type: string format: date-time description: >- The creation time of the deployment shape version. Lists will be ordered by this field. readOnly: true snapshot: $ref: "#/components/schemas/gatewayDeploymentShape" description: Full snapshot of the Deployment Shape at this version. readOnly: true validated: type: boolean description: If true, this version has been validated. public: type: boolean description: If true, this version will be publicly readable. latestValidated: type: boolean description: |- If true, this version is the latest validated version. Only one version of the shape can be the latest validated version. readOnly: true capabilities: type: array items: $ref: "#/components/schemas/DeploymentShapeVersionCapability" description: The capabilities supported by this deployment shape version. title: >- A deployment shape version is a specific version of a deployment shape. Versions are immutable, only created on updates and deleted when the deployment shape is deleted. gatewayDeploymentShard: type: object properties: name: type: string title: >- The resource name of the deployment shard. e.g. accounts/my-account/deployments/my-deployment/shards/my-deployment-shard readOnly: true createTime: type: string format: date-time description: The creation time of the deployment shard. readOnly: true updateTime: type: string format: date-time description: The update time for the deployment shard. readOnly: true state: $ref: "#/components/schemas/gatewayDeploymentShardState" description: The state of the deployment shard. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Detailed status information regarding the most recent operation. readOnly: true replicaStats: $ref: "#/components/schemas/gatewayReplicaStats" description: >- Per-replica deployment status counters for this shard. Provides visibility into pod states (pending, downloading, initializing, ready) across regions. readOnly: true gatewayDeploymentShardState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - DELETING - FAILED - UPDATING default: STATE_UNSPECIFIED description: |- - CREATING: The deployment shard is still being created. - READY: The deployment shard is ready to be used. - DELETING: The deployment shard is being deleted. - FAILED: The deployment shard failed to be created. See the `status` field for additional details on why it failed. - UPDATING: There are in-progress updates happening with the deployment shard. gatewayDeploymentState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - DELETING - FAILED - UPDATING - DELETED default: STATE_UNSPECIFIED description: |- - CREATING: The deployment is still being created. - READY: The deployment is ready to be used. - DELETING: The deployment is being deleted. - FAILED: The deployment failed to be created. See the `status` field for additional details on why it failed. - UPDATING: There are in-progress updates happening with the deployment. - DELETED: The deployment is soft-deleted. gatewayDeploymentTemplate: type: object properties: name: type: string title: >- The resource name of the template. e.g. accounts/my-account/deploymentTemplates/my-template readOnly: true displayName: type: string description: >- Human-readable display name of the deployment template. e.g. "My Template" Must be fewer than 64 characters long. description: type: string description: Description of the deployment template. createTime: type: string format: date-time description: The creation time of the deployment template. readOnly: true createdBy: type: string description: The email address of the user who created this deployment template. readOnly: true minReplicaCount: type: integer format: int32 maxReplicaCount: type: integer format: int32 autoscalingPolicy: $ref: "#/components/schemas/gatewayAutoscalingPolicy" baseModel: type: string acceleratorCount: type: integer format: int32 acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" enableAddons: type: boolean draftModel: type: string ngramSpeculationLength: type: integer format: int32 draftTokenCount: type: integer format: int32 enableSessionAffinity: type: boolean region: $ref: "#/components/schemas/gatewayRegion" directRouteApiKeys: type: array items: type: string directRouteType: $ref: "#/components/schemas/gatewayDirectRouteType" precision: $ref: "#/components/schemas/DeploymentPrecision" updateTime: type: string format: date-time description: The update time for the deployment template. readOnly: true maxContextLength: type: integer format: int32 description: >- The maximum context length supported by the model (context window). If set to 0 or not specified, the model's default maximum context length will be used. disableDeploymentSizeValidation: type: boolean description: Whether the deployment size validation is disabled. gatewayDeveloperPass: type: object properties: name: type: string title: |- The resource name of the developer pass. Format: accounts/{account_id}/developerPasses/{developer_pass_id} readOnly: true autoRenew: type: boolean description: Whether the developer pass will automatically renew upon expiry. endTime: type: string format: date-time description: The time at which the developer pass ends. readOnly: true createTime: type: string format: date-time description: Creation timestamp. readOnly: true updateTime: type: string format: date-time description: Last update timestamp. readOnly: true state: $ref: "#/components/schemas/gatewayDeveloperPassState" title: State of the developer pass readOnly: true statusMessage: type: string description: >- Additional context for the current state, such as a payment decline reason. readOnly: true lastRenewTime: type: string format: date-time description: The time at which the developer pass was last renewed. readOnly: true version: $ref: "#/components/schemas/DeveloperPassVersion" description: The version of developer pass (V1 = weekly, V2 = monthly). readOnly: true annotations: $ref: "#/components/schemas/gatewayDeveloperPassAnnotation" readOnly: true title: DeveloperPass represents an account-level usage pass gatewayDeveloperPassAnnotation: type: object properties: requestedVersion: $ref: "#/components/schemas/DeveloperPassVersion" description: Target version for a pending v1 -> v2 upgrade. promoCreditCode: type: string description: User-facing promo code that created this pass. gatewayDeveloperPassState: type: string enum: - STATE_UNSPECIFIED - ACTIVE - EXPIRED - FAILED - CREATING - ENABLING_AUTO_RENEW - DISABLING_AUTO_RENEW - PAYMENT_FAILED - UPDATING default: STATE_UNSPECIFIED gatewayDirectRouteType: type: string enum: - DIRECT_ROUTE_TYPE_UNSPECIFIED - INTERNET - GCP_PRIVATE_SERVICE_CONNECT - AWS_PRIVATELINK default: DIRECT_ROUTE_TYPE_UNSPECIFIED title: |- - DIRECT_ROUTE_TYPE_UNSPECIFIED: No direct routing - INTERNET: The direct route is exposed via the public internet - GCP_PRIVATE_SERVICE_CONNECT: The direct route is exposed via GCP Private Service Connect - AWS_PRIVATELINK: The direct route is exposed via AWS PrivateLink gatewayDistributionBucket: type: object properties: distributionBins: type: array items: type: number format: double description: Upper-edge values of the histogram bins. distributionCounts: type: array items: type: string format: int64 description: Per-bin sample counts; len == len(distribution_bins) - 1. average: type: number format: double median: type: number format: double successCount: type: string format: int64 description: Booleans like "format_valid" project into success/failure counts. failureCount: type: string format: int64 description: |- DistributionBucket holds a per-metric histogram + summary stats over a single chunk's worth of samples. gatewayDistributionSnapshot: type: object properties: metrics: type: object additionalProperties: $ref: "#/components/schemas/gatewayDistributionBucket" description: Keyed by metric name (e.g. "reward", "format_valid"). correlation: $ref: "#/components/schemas/gatewayCorrelationMatrix" gatewayDpoConfig: type: object properties: beta: type: number format: float description: >- DPO temperature parameter (beta in the paper). Must be > 0 and < 0.5. refCacheConcurrency: type: integer format: int32 description: Max concurrent reference forward passes during cache warm-up. refCacheBatchSize: type: integer format: int32 description: >- Number of preference pairs per reference forward call during caching. description: Hyperparameters for Direct Preference Optimization (DPO) training. gatewayDpoJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true completedTime: type: string format: date-time readOnly: true dataset: type: string description: The name of the dataset used for training. state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true createdBy: type: string description: The email address of the user who initiated this dpo job. readOnly: true trainingConfig: $ref: "#/components/schemas/gatewayBaseTrainingConfig" description: Common training configurations. wandbConfig: $ref: "#/components/schemas/gatewayWandbConfig" description: The Weights & Biases team/user account for logging job progress. trainerLogsSignedUrl: type: string description: |- The signed URL for the trainer logs file (stdout/stderr). Only populated if the account has trainer log reading enabled. readOnly: true lossConfig: $ref: "#/components/schemas/gatewayReinforcementLearningLossConfig" description: |- Loss configuration for the training job. If not specified, defaults to DPO loss. Set method to ORPO for ORPO training. awsS3Config: $ref: "#/components/schemas/gatewayAwsS3Config" description: The AWS configuration for S3 dataset access. azureBlobStorageConfig: $ref: "#/components/schemas/gatewayAzureBlobStorageConfig" description: The Azure configuration for Azure Blob Storage dataset access. purpose: $ref: "#/components/schemas/gatewayPurpose" description: Scheduling purpose for this job. useReservation: type: boolean description: >- Whether to try the account's reservation capacity before falling back to shared trainer capacity. required: - dataset gatewayEagleTrainingJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string description: The email address of the user who created this EAGLE training job. readOnly: true state: $ref: "#/components/schemas/gatewayJobState" description: State of the training job. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Information describing success or failure of the training job. readOnly: true inputDraftModel: type: string description: |- The name of a model with Kind=DRAFT_ADDON. Model architecture must be either Llama or Mixtral and compatible with the base model. Example: "accounts/example-account/models/example-draft-model". trainingDataset: type: string description: |- The name of the dataset to use for training the eagle model. Example: "accounts/example-account/datasets/example-dataset". baseModel: type: string description: |- The name of the base model to use for training the eagle model. Model architecture must be either Llama or Mixtral. Kind must be HF_BASE_MODEL. Example: "accounts/example-account/models/example-base-model". If not specified, will be inferred from source_deployment or hidden_states_gen_config.deployed_model when available. outputDraftModel: type: string description: |- The name of the output draft model that is generated as the output of this job. Example: "accounts/example-account/models/example-draft-model". epochsCount: type: number format: float description: |- Number of epochs to train for. If absent, it will be set automatically. learningRate: type: number format: float description: |- Learning rate for the EAGLE training job. If absent, it will be set automatically. wandbConfig: $ref: "#/components/schemas/gatewayWandbConfig" description: >- The Weights & Biases team/user account for logging training progress. updateTime: type: string format: date-time description: The update time for the eagle training job. readOnly: true hiddenStatesGenConfig: $ref: "#/components/schemas/gatewayHiddenStatesGenConfig" description: Configuration for hidden states generation during data preparation. intermediateSize: type: integer format: int32 description: Intermediate size for model architecture. numHiddenLayers: type: integer format: int32 description: Number of hidden layers for model architecture. maxContextLen: type: integer format: int32 description: The maximum context length to use with the model. batchSize: type: integer format: int32 description: >- The maximum number of tokens per batch for training in sequence packing. skipDataGen: type: boolean description: Whether to skip data generation and use existing data. dataGenOnly: type: boolean description: Whether to only perform data generation without training. trainingStyle: $ref: "#/components/schemas/EagleTrainingJobTrainingStyle" description: The training style for the EAGLE training job. gatewayEksCluster: type: object properties: awsAccountId: type: string description: The 12-digit AWS account ID where this cluster lives. fireworksManagerRole: type: string title: >- The IAM role ARN used to manage Fireworks resources on AWS. If not specified, the default is arn:aws:iam:::role/FireworksManagerRole region: type: string description: >- The AWS region where this cluster lives. See https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Concepts.RegionsAndAvailabilityZones.html for a list of available regions. clusterName: type: string description: The EKS cluster name. storageBucketName: type: string description: The S3 bucket name. metricWriterRole: type: string description: >- The IAM role ARN used by Google Managed Prometheus role that will write metrics to Fireworks managed Prometheus. The role must be assumable by the `system:serviceaccount:gmp-system:collector` service account on the EKS cluster. If not specified, no metrics will be written to GCP. loadBalancerControllerRole: type: string description: >- The IAM role ARN used by the EKS load balancer controller (i.e. the load balancer automatically created for the k8s gateway resource). If not specified, no gateway will be created. workloadIdentityPoolProviderId: type: string title: |- The ID of the GCP workload identity pool provider in the Fireworks project for this cluster. The pool ID is assumed to be "byoc-pool" inferenceRole: type: string description: The IAM role ARN used by the inference pods on the cluster. title: An Amazon Elastic Kubernetes Service cluster. required: - awsAccountId - region gatewayEncryptionState: type: string enum: - ENCRYPTION_STATE_UNSPECIFIED - ENCRYPTION_STATE_PLAINTEXT - ENCRYPTION_STATE_CMEK default: ENCRYPTION_STATE_UNSPECIFIED description: >- EncryptionState is the authoritative, per-resource CMEK marker: the single source of truth for whether a resource's customer-data artifacts are encrypted. It is stamped at resource creation from the owning account's CMEK config and is immutable for the resource's life, so a resource always reads back with the state it was created under regardless of the account's current flag. The encrypt-on-write / decrypt-on-read decision for the resource's data keys off this field. - ENCRYPTION_STATE_UNSPECIFIED: Unstamped/legacy rows that predate CMEK. Treated as PLAINTEXT. - ENCRYPTION_STATE_PLAINTEXT: Customer-data artifacts are stored as plaintext (Fireworks-owned, no CMEK). - ENCRYPTION_STATE_CMEK: Customer-data artifacts are CMEK-encrypted; reads MUST decrypt (fail-closed). gatewayEstimateSupervisedFineTuningJobCostResponse: type: object properties: estimatedCost: $ref: "#/components/schemas/typeMoney" title: An estimate of the fine-tuning costs gatewayEvalProtocol: type: object gatewayEvaluateOptions: type: object properties: maxConcurrency: type: integer format: int32 title: "Maximum concurrent requests (default: 4)" repeat: type: integer format: int32 title: "Number of times to repeat each test case (default: 1)" delay: type: integer format: int32 title: Delay (in ms) between API calls title: Options for how to run the evaluation gatewayEvaluation: type: object properties: name: type: string title: Current fields in your proto readOnly: true createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true evaluationType: type: string title: string llm_evaluator_prompt = 6; description: type: string title: Optional description of the evaluation providers: type: array items: type: object $ref: "#/components/schemas/gatewayProvider" title: One or more providers to use assertions: type: array items: type: object $ref: "#/components/schemas/gatewayAssertion" title: One or more assertions to evaluate updateTime: type: string format: date-time description: The update time for the evaluation. readOnly: true required: - evaluationType - providers - assertions gatewayEvaluationJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true evaluator: type: string description: >- The fully-qualified resource name of the Evaluation used by this job. Format: accounts/{account_id}/evaluators/{evaluator_id} inputDataset: type: string description: >- The fully-qualified resource name of the input Dataset used by this job. Format: accounts/{account_id}/datasets/{dataset_id} outputDataset: type: string description: >- The fully-qualified resource name of the output Dataset created by this job. Format: accounts/{account_id}/datasets/{output_dataset_id} metrics: type: object additionalProperties: type: number format: double readOnly: true outputStats: type: string description: The output dataset's aggregated stats for the evaluation job. updateTime: type: string format: date-time description: The update time for the evaluation job. readOnly: true awsS3Config: $ref: "#/components/schemas/gatewayAwsS3Config" description: The AWS configuration for S3 dataset access. required: - evaluator - inputDataset - outputDataset gatewayEvaluationResult: type: object properties: evaluationJobId: type: string required: - evaluationJobId gatewayEvaluator: type: object properties: name: type: string readOnly: true displayName: type: string description: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true updateTime: type: string format: date-time readOnly: true state: $ref: "#/components/schemas/gatewayEvaluatorState" readOnly: true criteria: type: array items: type: object $ref: "#/components/schemas/gatewayCriterion" title: >- Criteria for the evaluator, it should produce a score for the metric (name of criteria) Used for eval3 with UI upload path requirements: type: string title: Content for the requirements.txt for package installation entryPoint: type: string title: >- entry point of the evaluator inside the codebase. In module::function or path::function format status: $ref: "#/components/schemas/gatewayStatus" title: Status of the evaluator, used to expose build status to the user readOnly: true commitHash: type: string title: Commit hash of this evaluator from the user's original codebase source: $ref: "#/components/schemas/EvaluatorSource" description: Source information for the evaluator codebase. defaultDataset: type: string title: Default dataset that is associated with the evaluator encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true gatewayEvaluatorState: type: string enum: - STATE_UNSPECIFIED - ACTIVE - BUILDING - BUILD_FAILED default: STATE_UNSPECIFIED title: |- - ACTIVE: The evaluator is ready to use for evaluation - BUILDING: The evaluator is being built, i.e. building the e2b template - BUILD_FAILED: The evaluator build failed, and it cannot be used for evaluation gatewayEvaluatorVersion: type: object properties: name: type: string description: The name of the evaluator revision. readOnly: true snapshot: $ref: "#/components/schemas/gatewayEvaluator" description: The snapshot of the evaluator at the time the revision was created. readOnly: true createTime: type: string format: date-time description: The timestamp that the revision was created. readOnly: true alternateIds: type: array items: type: string description: >- Other revision IDs that share the same snapshot (e.g. aliases like "latest"). readOnly: true gatewayEvenLoadStrategy: type: object description: >- Dynamically adjust traffic allocation to balance the load per replica across the deployments as much as possible. gatewayExample: type: object properties: content: type: string gatewayExportBillingMetricsResponse: type: object properties: signedUrls: type: array items: type: string description: |- The signed URL of the exported file(s). There will be exactly one file. This may change in the future. gatewayExternalKey: type: object properties: name: type: string title: |- The resource name of the external key. Format: accounts/{account}/externalKeys/{external_key} readOnly: true displayName: type: string description: Display name for the external key. createTime: type: string format: date-time description: Creation timestamp. readOnly: true updateTime: type: string format: date-time description: Last update timestamp. readOnly: true state: $ref: "#/components/schemas/gatewayExternalKeyState" description: Current state of the external key. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the external key status. readOnly: true description: >- ExternalKey is an account-level, customer-managed KMS key registration for CMEK envelope encryption. It is a collection — an account may hold several keys at once — with exactly one marked primary. The primary wraps newly minted DEKs; older keys stay active so data wrapped under them still decrypts (the rotation/migration story is inherently multi-key). The whole resource is Fireworks-operator (superuser) managed. Fireworks reaches the customer KMS as a single shared Google OIDC identity (a global per-env subject), scoped per account by minting tokens with aud = account_id, pinned in the customer's own KMS trust. So these configs carry identity *references* only — never long-lived secrets — and Fireworks provisions nothing per account. gatewayExternalKeyState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - UPDATING - DELETING default: STATE_UNSPECIFIED description: State of the external key. gatewayFakeCluster: type: object properties: projectId: type: string location: type: string clusterName: type: string title: A fake cluster using https://pkg.go.dev/k8s.io/client-go/kubernetes/fake gatewayFeatureFlag: type: object properties: name: type: string title: The resource name, e.g. accounts/my-account/featureFlags/my-feature readOnly: true value: type: string description: |- Additional human-readable feature value if needed. If not set, the default value is "true". createTime: type: string format: date-time readOnly: true gatewayGatewayAbortTrainingRunMaterializationBody: type: object properties: status: $ref: "#/components/schemas/gatewayStatus" description: |- Failure reason to persist on the run. If omitted, the server uses UNAVAILABLE with a generic materialization failure message. gatewayGatewayAliasEvaluatorVersionBody: type: object properties: aliasId: type: string description: The alias ID (e.g., "current", a semver tag, etc.). description: Assigns or updates a user-specified alias for a given revision. required: - aliasId gatewayGatewayApproveUserUsageLimitIncreaseRequestBody: type: object gatewayGatewayCancelDpoJobBody: type: object gatewayGatewayCancelReinforcementFineTuningJobBody: type: object gatewayGatewayCancelRlorTrainerJobBody: type: object gatewayGatewayCancelSupervisedFineTuningJobBody: type: object gatewayGatewayCreateApiKeyBody: type: object properties: apiKey: $ref: "#/components/schemas/gatewayApiKey" description: The API key to be created. required: - apiKey gatewayGatewayCreateClusterBody: type: object properties: cluster: $ref: "#/components/schemas/gatewayCluster" description: The properties of the cluster being created. clusterId: type: string title: The cluster ID to use in the cluster name. e.g. my-cluster required: - cluster - clusterId gatewayGatewayCreateDatasetBody: type: object properties: dataset: $ref: "#/components/schemas/gatewayDataset" datasetId: type: string sourceDatasetId: type: string title: >- If set, indicates we are creating a new dataset by filtering this existing dataset ID filter: type: string title: >- Filter condition (SQL-like WHERE clause) to apply to the source dataset required: - dataset - datasetId gatewayGatewayCreateEvaluationBody: type: object properties: evaluation: $ref: "#/components/schemas/gatewayEvaluation" evaluationId: type: string required: - evaluation gatewayGatewayCreateEvaluationJobBody: type: object properties: evaluationJob: $ref: "#/components/schemas/gatewayEvaluationJob" evaluationJobId: type: string required: - evaluationJob gatewayGatewayCreateEvaluatorBody: type: object properties: evaluator: $ref: "#/components/schemas/gatewayEvaluator" evaluatorId: type: string required: - evaluator gatewayGatewayCreateEvaluatorV2Body: type: object properties: evaluator: $ref: "#/components/schemas/gatewayEvaluator" evaluatorId: type: string required: - evaluator gatewayGatewayCreateModelBody: type: object properties: model: $ref: "#/components/schemas/gatewayModel" description: The properties of the Model being created. modelId: type: string description: ID of the model. cluster: type: string description: |- The resource name of the BYOC cluster to which this model belongs. e.g. accounts/my-account/clusters/my-cluster. Empty if it belongs to a Fireworks cluster. required: - modelId gatewayGatewayDebugReinforcementFineTuningJobBody: type: object gatewayGatewayDeleteApiKeyBody: type: object properties: keyId: type: string description: The key ID for the API key. required: - keyId gatewayGatewayDeployModelVersionBody: type: object properties: deployment: type: string title: The deployment name to deploy the model version to rolloutStrategy: $ref: "#/components/schemas/gatewayRolloutStrategy" title: The rollout strategy to use when deploying the model version required: - deployment - rolloutStrategy gatewayGatewayFinalizeTrainingRunMaterializationBody: type: object properties: modelId: type: string description: >- Trainer-returned model id. Must be run-scoped, e.g. "{run_id}:train:0". required: - modelId gatewayGatewayGetDatasetUploadEndpointBody: type: object properties: filenameToSize: type: object additionalProperties: type: string format: int64 description: A mapping from the file name to its size in bytes. readMask: type: string description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. required: - filenameToSize gatewayGatewayGetEvaluatorUploadEndpointBody: type: object properties: filenameToSize: type: object additionalProperties: type: string format: int64 readMask: type: string required: - filenameToSize gatewayGatewayGetModelUploadEndpointBody: type: object properties: filenameToSize: type: object additionalProperties: type: string format: int64 description: A mapping from the file name to its size in bytes. enableResumableUpload: type: boolean description: If true, enable resumable upload instead of PUT. readMask: type: string description: >- The fields to be returned in the response. If empty or "*", all fields will be returned. required: - filenameToSize gatewayGatewayImportModelBody: type: object properties: awsS3Source: $ref: "#/components/schemas/gatewayAwsS3ModelSource" description: |- AWS S3 source details. Must be set when importing from AWS S3. azureBlobSource: $ref: "#/components/schemas/gatewayAzureBlobModelSource" description: |- Azure Blob Storage source details. Must be set when importing from Azure Blob Storage. gatewayGatewayLoadTrainingSessionStateBody: type: object properties: path: type: string description: Adapter checkpoint path to load into the training session. description: |- LoadTrainingSessionState loads/switches the LoRA adapter on a session. The `Execute*` prefix is the Fireworks-internal convention for side-effecting state transitions on an existing resource (vs. AIP-136's `LoadStateTrainingSession`); we keep it for repo-wide consistency. required: - path gatewayGatewayLogTrainingSessionMetricsBody: type: object properties: points: type: array items: type: object $ref: "#/components/schemas/gatewayMetricPoint" description: >- Batch of metric observations. Server caps the per-request batch size to bound upsert statement parameter count. required: - points gatewayGatewayPrepareModelBody: type: object properties: precision: $ref: "#/components/schemas/DeploymentPrecision" title: the precision with which the model will be prepared readMask: type: string title: >- The fields to be returned in the response. If empty or "*", all fields will be returned. This is added as is used in getResource() acceleratorCount: type: integer format: int32 description: >- Number of accelerators (GPUs) to use for quantization. Defaults to 8 if unset. abort: type: boolean description: >- If true, abort an active prepare job and clear prepare tracking state. No new prepare job is launched in this request. gatewayGatewayPreviewEvaluationBody: type: object properties: sampleData: type: string title: Sample data in JSON format (array of samples) maxSamples: type: integer format: int32 title: >- Maximum number of samples to evaluate (optional, default will be a small number like 5) title: Request to preview an evaluation with sample data required: - sampleData gatewayGatewayPreviewEvaluatorBody: type: object properties: evaluator: $ref: "#/components/schemas/gatewayEvaluator" title: The evaluator object to run preview on sampleData: type: array items: type: string title: Sample data in JSON format (array of json formated samples) maxSamples: type: integer format: int32 title: Maximum number of samples to evaluate, default is 10 required: - evaluator - sampleData gatewayGatewayPromoteCheckpointBody: type: object properties: outputModel: type: string title: |- The desired output model resource name. Format: accounts/{account}/models/{model} trainerJobId: type: string title: >- Deprecated. Only used by legacy callers passing the flat `name` form; new callers pass the nested `name` (trainer id embedded) and leave this unset. If provided, must agree with the trainer id in `name`. Format: accounts/{account}/rlorTrainerJobs/{rlor_trainer_job} baseModel: type: string title: >- The base model for metadata inheritance (model type, context length, etc.). Format: accounts/{account}/models/{model} hotLoadDeploymentId: type: string title: >- Deprecated. Callers should not set this — the server resolves the owning GCS path from the trainer job's stored hot_load_bucket_url (the same source the list endpoint uses). Silently ignored when the trainer row is present. Retained for request-shape compatibility with old clients. Format: accounts/{account}/deployments/{deployment} asyncPromotion: type: boolean description: >- Optional request-level opt-in for long-running checkpoint promotion. When false or unset, the server preserves the legacy synchronous response shape and returns `model`. When true, the server may return `operation`; callers must poll GetOperation until completion and read the promoted model from the operation response. description: >- Promote a checkpoint to a model. Pass the `name` returned by ListRlorTrainerJobCheckpoints verbatim; the server parses the trainer id out of the resource name and reads the trainer's stored hot_load_bucket_url — the same authoritative source the list endpoint uses — to locate the checkpoint in GCS. Callers do NOT need to pass `trainer_job_id` or `hot_load_deployment_id` when `name` is the nested form. required: - outputModel - baseModel gatewayGatewayPromoteTrainingSessionCheckpointBody: type: object properties: outputModel: type: string description: >- Output Model resource name (accounts/{a}/models/{m}); the account must match the checkpoint's account. baseModel: type: string description: Base Model resource name the training run started from. required: - outputModel - baseModel gatewayGatewayRejectUserUsageLimitIncreaseRequestBody: type: object gatewayGatewayResumeDpoJobBody: type: object gatewayGatewayResumeReinforcementFineTuningJobBody: type: object gatewayGatewayResumeRlorTrainerJobBody: type: object gatewayGatewayResumeSupervisedFineTuningJobBody: type: object gatewayGatewayRollbackEvaluatorBody: type: object description: >- Rolls an evaluator back to the configuration captured by a specific revision. gatewayGatewayScaleDeploymentBody: type: object properties: replicaCount: type: integer format: int32 description: The desired number of replicas. gatewayGatewaySplitDatasetBody: type: object properties: chunkSize: type: integer format: int32 title: Required. The size of each chunk (minimum 200) parent: type: string description: The parent account ID of the requester. title: Request message for splitting a dataset into chunks gatewayGatewaySyncRlorTrainerJobCheckpointBody: type: object properties: region: $ref: "#/components/schemas/gatewayRegion" description: >- Runtime region for the SINGLE_REGION side of the sync. If target_availability is MANAGED_DURABLE, this is the source region that already contains the checkpoint. If target_availability is SINGLE_REGION, this is the destination region that should receive a runtime-local copy. If unset, the trainer job's configured region is used. targetAvailability: $ref: "#/components/schemas/CheckpointAvailabilityAvailabilityClass" description: >- Target storage availability to ensure. If unset or AVAILABILITY_CLASS_UNSPECIFIED, the server treats this as MANAGED_DURABLE: it ensures a durable managed copy exists, using `region` as the source runtime region or the trainer job's configured region when `region` is also unset. Set SINGLE_REGION to explicitly stage a durable checkpoint into `region` before restoring there. The operation is idempotent: if the requested target availability already exists, the server returns the checkpoint with its current availability. gatewayGatewayTestEvaluationBody: type: object properties: evaluation: $ref: "#/components/schemas/gatewayEvaluation" title: The draft evaluation to test sampleData: type: string title: Sample data in JSON format (array of samples) required: - evaluation - sampleData gatewayGatewayUndeleteDeploymentBody: type: object gatewayGatewayValidateAssertionsBody: type: object properties: assertions: type: array items: type: object $ref: "#/components/schemas/gatewayAssertion" description: >- Validate the code assertions. Recommended before creating the evaluation. required: - assertions gatewayGatewayValidateDatasetUploadBody: type: object gatewayGatewayValidateEvaluatorUploadBody: type: object gatewayGetAccountServerlessTokenUsageResponse: type: object properties: averageTokensPerMinuteByBaseModel: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" totalPeakGeneratedTokensPerMinute: $ref: "#/components/schemas/gatewayTimeSeries" description: Account-wide peak TPM by token category. totalPeakUncachedPromptTokensPerMinute: $ref: "#/components/schemas/gatewayTimeSeries" totalPeakCachedPromptTokensPerMinute: $ref: "#/components/schemas/gatewayTimeSeries" peakGeneratedTokensPerMinuteByBaseModel: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" description: Peak TPM per base model and token category. peakUncachedPromptTokensPerMinuteByBaseModel: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" peakCachedPromptTokensPerMinuteByBaseModel: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" peakRejectedPromptTokensPerMinuteByBaseModel: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" gatewayGetAccountUsageFilterOptionsRequestUsageType: type: string enum: - USAGE_TYPE_UNSPECIFIED - SERVERLESS - DEDICATED_DEPLOYMENT - TRAINING default: USAGE_TYPE_UNSPECIFIED description: >- Which usage stream(s) to load distinct filter values for. Mirrors GetAccountUsageRequest.UsageType. UNSPECIFIED preserves the legacy behavior: load filter options for serverless and dedicated usage only. - SERVERLESS: Serverless filter dimensions only (model_name, api_key_id, api_key_name, user_id, session_id, annotations.*). - DEDICATED_DEPLOYMENT: Dedicated deployment filter dimensions (deployment_name, annotations.team, .project, .environment). - TRAINING: Training filter dimensions (job_id, training_session_id, job_type, usage_type, accelerator_type, base_model). gatewayGetAccountUsageFilterOptionsResponse: type: object description: >- Distinct filter values for the account/time range (separate from GetAccountUsage). gatewayGetAccountUsageRequestUsageType: type: string enum: - USAGE_TYPE_UNSPECIFIED - SERVERLESS - DEDICATED_DEPLOYMENT - TRAINING default: USAGE_TYPE_UNSPECIFIED description: >- - USAGE_TYPE_UNSPECIFIED: Default value. When specified (or when usage_type field is not set), returns usage data for all deployment types: serverless requests, dedicated deployments, and training jobs. - SERVERLESS: Returns only serverless usage data. Filters the response to include only usage from serverless API requests. - DEDICATED_DEPLOYMENT: Returns only dedicated deployment usage data. Filters the response to include only usage from dedicated deployments. - TRAINING: Returns only training job usage data (SFT/DPO token usage and RFT / service-mode trainer GPU-seconds usage). Inference deployments serving rollouts for RFT / online RL are reported under DEDICATED_DEPLOYMENT (not TRAINING) to avoid double counting GPU time. title: Usage type to query usage for gatewayGetBatchInferenceJobInputUploadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string description: Signed URLs for users to upload their input to GCS. gatewayGetBatchInferenceJobOutputDownloadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string description: Signed URLs for users to download their input from GCS. gatewayGatewayQueryUsageCostsBody: type: object properties: startTime: type: string format: date-time description: Inclusive lower bound for usage event timestamps. endTime: type: string format: date-time description: Exclusive upper bound for usage event timestamps. scope: $ref: "#/components/schemas/gatewayQueryUsageCostsRequestScope" description: >- ACCOUNT requires account administrator access. SELF derives the user ID from the authenticated principal. groupBy: type: array items: $ref: "#/components/schemas/QueryUsageCostsRequestDimension" description: >- Dimensions to group by. At most two dimensions may be combined. HOUR and DAY are mutually exclusive. SESSION is intentionally filter-only to prevent high-cardinality enumeration. filter: $ref: "#/components/schemas/gatewayUsageCostFilter" description: Optional exact-match filters. All populated fields are ANDed. pageSize: type: integer format: int32 description: >- For paginated queries, the maximum number of grouped values to return. A page can contain more rows when each value spans multiple time buckets. Zero uses the server default; values above the maximum are clamped. pageToken: type: string description: >- Opaque token from a previous response. Set this to retrieve the next page while keeping all other request fields unchanged. maxLength: 4096 description: Request for grouped usage cost subtotals over a time range. gatewayQueryUsageCostsRequestScope: type: string enum: - SCOPE_UNSPECIFIED - ACCOUNT - SELF default: SCOPE_UNSPECIFIED gatewayQueryUsageCostsResponse: type: object properties: rows: type: array items: type: object $ref: "#/components/schemas/gatewayUsageCostRow" description: >- Cost rows for the requested dimensions. A successful response contains at most 10,100 rows and is limited to 4 MiB. nextPageToken: type: string description: Opaque token for the next page. Empty when no additional page is available. subtotal: $ref: "#/components/schemas/typeMoney" description: Subtotal for the complete query. evaluationTime: type: string format: date-time description: Time when this result was evaluated. readOnly: true attributionCompleteness: $ref: "#/components/schemas/QueryUsageCostsResponseAttributionCompleteness" description: >- Whether the result may omit usage without a user ID. INCOMPLETE means the query filters by user, so historical unattributed usage may be excluded. The amount excluded is unknown. readOnly: true QueryUsageCostsRequestDimension: type: string enum: - DIMENSION_UNSPECIFIED - HOUR - MODEL - USER - API_KEY - DAY default: DIMENSION_UNSPECIFIED QueryUsageCostsResponseAttributionCompleteness: type: string enum: - ATTRIBUTION_COMPLETENESS_UNSPECIFIED - COMPLETE - INCOMPLETE - UNKNOWN default: ATTRIBUTION_COMPLETENESS_UNSPECIFIED gatewayUsageCostDimensions: type: object properties: startTime: type: string format: date-time description: >- Start of the UTC HOUR or DAY bucket intersected with the requested range. The first bucket may start at the request's start_time. model: type: string description: Canonical model resource name. user: type: string description: User resource name. apiKeyId: type: string description: Stable API key ID. unattributedDimensions: type: array items: $ref: "#/components/schemas/QueryUsageCostsRequestDimension" description: >- Requested dimensions that were absent from the billing event. More than one dimension may be absent when the query groups by two dimensions. readOnly: true unknownModel: type: boolean description: >- Whether a populated billing model could not be mapped to a public model. The internal model value is not exposed. readOnly: true description: >- Grouping values for one usage cost row. Fields not requested in group_by are left empty. gatewayUsageCostFilter: type: object properties: model: type: string title: |- Canonical model resource name. Format: accounts/{account}/models/{model} user: type: string title: |- User resource name. Valid only for account-scoped queries. Format: accounts/{account}/users/{user} apiKeyId: type: string description: Stable API key ID. This is never plaintext key material. sessionId: type: string description: >- Exact client session ID. Valid only for self-scoped queries. This value is used only as a filter and is never returned in the response. description: Exact property filters for a usage cost query. Populated fields are ANDed. gatewayUsageCostRow: type: object properties: dimensions: $ref: "#/components/schemas/gatewayUsageCostDimensions" subtotal: $ref: "#/components/schemas/typeMoney" description: >- Usage priced with the applicable subscription prices. This excludes fixed fees and invoice-level discounts, minimums, credits, and taxes. description: One row in a usage cost query. gatewayGetBillingSummaryResponse: type: object properties: lineItems: type: array items: type: object $ref: "#/components/schemas/gatewayLineItem" title: Individual billing line items usageBuckets: type: array items: type: object $ref: "#/components/schemas/gatewayUsageBucket" description: >- Per-bucket breakdown, populated only when the request sets a non-UNSPECIFIED granularity. One bucket per period at that granularity. title: Response for billing summary endpoint gatewayGetDatasetDownloadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string title: Signed URLs for downloading dataset files gatewayGetDatasetUploadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string title: Signed URLs for uploading dataset files gatewayGetDeploymentMetricsResponse: type: object properties: metrics: type: object additionalProperties: type: number format: double title: Map of metric name to metric value for the specified time range gatewayGetDpoJobMetricsFileResponse: type: object properties: signedUrl: type: string title: The signed URL for the metrics file title: |- when the JobMetrics file has been created for the DPO job and the file exists, we will populate this field empty otherwise gatewayGetEvaluationJobExecutionLogEndpointResponse: type: object properties: executionLogSignedUri: type: string description: >- Short-lived signed URL for the execution log file. Empty if the log file has not been created yet (e.g. job not started or still initializing). contentType: type: string description: |- Content type for the log file (e.g. "text/plain"). Only set when execution_log_signed_uri is present. expireTime: type: string format: date-time description: |- Expiration time of the signed URL. Only set when execution_log_signed_uri is present. description: Response carries the stream log URL (for VirtualizedLogViewer). gatewayGetEvaluatorBuildLogEndpointResponse: type: object properties: buildLogSignedUri: type: string title: Signed URL for the build log gatewayGetEvaluatorSourceCodeEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string title: Mapping from filename to signed URL for downloading the source code gatewayGetEvaluatorUploadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string gatewayGetLedgerResponse: type: object properties: ledger: type: array items: type: object $ref: "#/components/schemas/gatewayLedgerEntry" description: The contents of the ledger. gatewayGetModelDownloadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string title: Signed URLs for for downloading model files gatewayGetModelUploadEndpointResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string title: Signed URLs for uploading model files filenameToUnsignedUris: type: object additionalProperties: type: string description: >- Unsigned URIs (e.g. s3://bucket/key, gs://bucket/key) for uploading model files. Returned when the caller has permission to upload to the URIs. gatewayGetModelVersionCountResponse: type: object properties: count: type: integer format: int32 gatewayGetOAuthArgumentsResponse: type: object properties: issuerUrl: type: string clientId: type: string cognitoDomain: type: string gatewayGetReinforcementFineTuningJobMetricsFileResponse: type: object properties: signedUrl: type: string description: The signed URL for the metrics file. gatewayGetReservationDataResponse: type: object properties: reservationDataByType: type: array items: type: object $ref: >- #/components/schemas/GetReservationDataResponseReservationDataByType title: GPU counts broken down by accelerator type title: Response with current GPU usage and reservation counts gatewayGetRlorTrainerJobPublicLogsResponse: type: object properties: signedUrl: type: string description: >- Signed URL for the public trainer logs file (human-readable status updates). The URL expires after 24 hours and should be fetched dynamically when needed. gatewayGetTerminationMessageResponse: type: object properties: message: type: string description: The termination message. gatewayGetTotalHistoricalSpendResponse: type: object properties: spend: $ref: "#/components/schemas/typeMoney" gatewayHiddenStatesGenConfig: type: object properties: deployedModel: type: string description: >- The deployed model or deployment to use for hidden states generation. Can be either a DeployedModel resource name (accounts/{account}/deployedModels/{id}) or a Deployment resource name (accounts/{account}/deployments/{id}). When empty and deployment_shape is set, a deployment will be auto-created. maxWorkers: type: integer format: int32 maxTokens: type: integer format: int32 inputOffset: type: integer format: int32 inputLimit: type: integer format: int32 regenerateAssistant: type: boolean outputActivations: type: boolean apiKey: type: string deploymentShape: type: string description: >- Deployment shape to use when auto-creating a deployment for hidden states generation. When set and deployed_model is empty, the system will automatically create a deployment with this shape, wait for it to become ready, and clean it up after the job completes. replicaCount: type: integer format: int32 description: Number of replicas for the auto-created deployment. Defaults to 1. toleration: type: string description: >- Taint key to tolerate for the auto-created deployment. When unset, the deployment uses the default placement behavior. title: >- Config for generating dataset with hidden states for SFTJ or eagle training. gatewayIdentityProvider: type: object properties: name: type: string title: |- The resource name of the identity provider. Format: accounts/{account}/identityProviders/{identity_provider} readOnly: true displayName: type: string title: Display name for the identity provider createTime: type: string format: date-time title: Creation timestamp readOnly: true updateTime: type: string format: date-time title: Last update timestamp readOnly: true samlConfig: $ref: "#/components/schemas/gatewaySamlConfig" oidcConfig: $ref: "#/components/schemas/gatewayOidcConfig" tenantDomains: type: array items: type: string title: |- List of allowed domains for this identity provider Example: ["example.com", "example.co.uk", "example.de"] If not provided, domain will be derived from the account email state: $ref: "#/components/schemas/gatewayIdentityProviderState" title: Current state of the identity provider readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the identity provider status. readOnly: true domainUrl: type: string description: The domain URL. readOnly: true issuerUrl: type: string description: The OIDC issuer URL. readOnly: true clientId: type: string description: The OIDC client ID. readOnly: true enableJitUserProvisioning: type: boolean description: >- Enable Just-In-Time (JIT) user provisioning. When enabled, users are automatically created in the account on first SSO login if they don't already exist. When disabled, users must be pre-provisioned before they can authenticate via SSO. jitDefaultRole: type: string description: >- Default role assigned to JIT-provisioned users. Valid values: "admin", "user", "contributor", "inference-user". Only applies when enable_jit_user_provisioning is true and RBAC V2 is enabled. If empty or unset, defaults to "inference-user" (least privilege). If RBAC V2 is not enabled for the account, JIT users always get "user" role. enforceSso: type: boolean title: >- Enforce SSO authentication and restrict account access to users with approved email domains. When enabled: - Users with email domains matching tenant_domains must authenticate via this identity provider - Users with other email domains are blocked (tenant_domains becomes an allowlist) - Superusers and API keys are exempt from this enforcement enableIdpInitiatedSso: type: boolean description: >- Enable IdP-initiated SAML (Security Assertion Markup Language) single sign-on. When enabled, users can start the login flow from their identity provider's portal (e.g., Okta app launcher) instead of from the Fireworks login page. Only supported for SAML identity providers. gatewayIdentityProviderState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - UPDATING - DELETING default: STATE_UNSPECIFIED title: State of the identity provider gatewayInferenceLog: type: object properties: name: type: string title: >- Resource name in the format "accounts/{account_id}/inference-logs/{inference_log_id}" readOnly: true createTime: type: string format: date-time title: The time the inference log was created readOnly: true model: type: string title: The model used for the inference readOnly: true requestType: type: string title: Request type (chat, completion, embedding, etc.) readOnly: true inputContent: type: string title: The input content provided in the request readOnly: true outputContent: type: string title: The output content generated by the model readOnly: true durationMs: type: string format: int64 title: Duration of the request in milliseconds readOnly: true statusCode: type: integer format: int32 title: Status code of the API call readOnly: true metadata: type: object additionalProperties: type: string title: Additional metadata about the log readOnly: true updateTime: type: string format: date-time title: The update time for the inference log readOnly: true title: Trace resource definition gatewayInvoice: type: object properties: id: type: string title: ID of the invoice amountDue: $ref: "#/components/schemas/typeMoney" description: |- This is the final amount required to be charged to the customer after any credits, discounts and customer balance have been applied. invoiceUrl: type: string description: A URL for the invoice portal. state: $ref: "#/components/schemas/gatewayInvoiceState" targetTime: type: string format: date-time description: The scheduled time of the invoice. paidTime: type: string format: date-time description: The timestamp when the invoice was paid. type: $ref: "#/components/schemas/gatewayInvoiceType" title: The invoice type autoCollectionEnabled: type: boolean description: Whether Orb payment auto-collection is enabled for this invoice. gatewayInvoiceState: type: string enum: - STATE_UNSPECIFIED - DRAFT - ISSUED - PAID - VOID - FAILED default: STATE_UNSPECIFIED gatewayInvoiceType: type: string enum: - TYPE_UNSPECIFIED - POSTPAID_BILLING - PREPAID_CREDITS - CONTRACTED default: TYPE_UNSPECIFIED gatewayJobProgress: type: object properties: percent: type: integer format: int32 description: Progress percent, within the range from 0 to 100. epoch: type: integer format: int32 description: >- The epoch for which the progress percent is reported, usually starting from 0. This is optional for jobs that don't run in an epoch fasion, e.g. BIJ, EVJ. totalInputRequests: type: integer format: int32 description: Total number of input requests/rows in the job. totalProcessedRequests: type: integer format: int32 description: >- Total number of requests that have been processed (successfully or failed). successfullyProcessedRequests: type: integer format: int32 description: Number of requests that were processed successfully. failedRequests: type: integer format: int32 description: Number of requests that failed to process. outputRows: type: integer format: int32 description: Number of output rows generated. inputTokens: type: integer format: int32 description: Total number of input tokens processed. outputTokens: type: integer format: int32 description: Total number of output tokens generated. cachedInputTokenCount: type: integer format: int32 description: The number of input tokens that hit the prompt cache. description: Progress of a job, e.g. RLOR, EVJ, BIJ etc. gatewayJobState: type: string enum: - JOB_STATE_UNSPECIFIED - JOB_STATE_CREATING - JOB_STATE_RUNNING - JOB_STATE_COMPLETED - JOB_STATE_FAILED - JOB_STATE_CANCELLED - JOB_STATE_DELETING - JOB_STATE_WRITING_RESULTS - JOB_STATE_VALIDATING - JOB_STATE_DELETING_CLEANING_UP - JOB_STATE_PENDING - JOB_STATE_EXPIRED - JOB_STATE_RE_QUEUEING - JOB_STATE_CREATING_INPUT_DATASET - JOB_STATE_IDLE - JOB_STATE_CANCELLING - JOB_STATE_EARLY_STOPPED - JOB_STATE_PAUSED - JOB_STATE_DELETED - JOB_STATE_ARCHIVED default: JOB_STATE_UNSPECIFIED description: |- JobState represents the state an asynchronous job can be in. - JOB_STATE_PAUSED: Job is paused, typically due to account suspension or manual intervention. - JOB_STATE_DELETED: Job has been deleted. - JOB_STATE_ARCHIVED: User-facing state for jobs whose row is retained post-delete (e.g. RLOR trainers within the checkpoint retention window). The internal row is still in JOB_STATE_DELETED; the gateway translates it to ARCHIVED on public responses. gatewayLLMAssertion: type: object properties: llmEvaluatorPrompt: type: string title: Prompt used to evaluate the output providers: type: array items: type: object $ref: "#/components/schemas/gatewayProvider" title: One or more providers to use prompts: type: array items: type: string title: One or more prompts to evaluate evaluateOptions: $ref: "#/components/schemas/gatewayEvaluateOptions" title: Options for how to run the evaluation required: - providers - prompts gatewayLearningRateScheduler: type: object properties: constant: $ref: "#/components/schemas/gatewayConstantLRSchedule" linear: $ref: "#/components/schemas/gatewayLinearLRSchedule" cosine: $ref: "#/components/schemas/gatewayCosineLRSchedule" description: >- Learning-rate scheduler applied on top of the base learning rate (after the shared learning_rate_warmup_steps warmup). It mirrors the trainer-side fireworks.training.sdk.training_spec.LRSchedulerSpec discriminated union: the schedule variant that is set selects the shape, and each variant carries only the knobs that apply to it. Add a new schedule by adding a message and one `oneof schedule` entry; existing clients are unaffected. When unset, the trainer uses the legacy constant schedule. gatewayLedgerEntry: type: object properties: timestamp: type: string format: date-time description: The timestamp of the entry. value: type: string description: The contents of the entry. gatewayLineItem: type: object properties: category: type: string title: Category name (e.g., "Audio Transcription", "Text Completion") groupingKey: type: string title: Primary grouping key (e.g., "model", "model_bucket") groupingValue: type: string title: >- Primary grouping value (e.g., "whisper-v3", "llama-v3p3-70b-instruct") secondaryGroupingKey: type: string title: Secondary grouping key (e.g., "is_batch") secondaryGroupingValue: type: string title: Secondary grouping value (e.g., "true", "false") quantity: type: number format: double title: Quantity consumed unitAmount: $ref: "#/components/schemas/typeMoney" title: Unit amount (price per unit) totalCost: $ref: "#/components/schemas/typeMoney" title: Total cost for this line item series: $ref: "#/components/schemas/gatewayGetAccountUsageRequestUsageType" description: >- Product-line series for this line item (serverless / dedicated / training). Defaults to USAGE_TYPE_UNSPECIFIED when not classified. title: Individual billing line item gatewayLinearLRSchedule: type: object properties: minLrRatio: type: number format: float description: >- Floor learning rate as a fraction of the peak learning rate (0.0 = decay to zero, 0.1 = decay to 10% of the peak learning rate). decayRatio: type: number format: float description: >- Fraction of total training steps over which to decay. 0 (unset) decays over the full run. description: >- Linear decay from the peak learning rate toward min_lr_ratio after warmup. gatewayListAccountServerlessRateLimitsResponse: type: object properties: rateLimits: type: array items: type: object $ref: "#/components/schemas/gatewayAccountRateLimit" description: >- Snapshot of effective limits at the end of the requested range (or at the current time if `end` is unset), using the same resolution rules as the time series. series: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" description: >- Time series of peak effective limits per interval, one series per deployment and metric. gatewayListAccountsResponse: type: object properties: accounts: type: array items: type: object $ref: "#/components/schemas/gatewayAccount" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of accounts. gatewayListApiKeysResponse: type: object properties: apiKeys: type: array items: type: object $ref: "#/components/schemas/gatewayApiKey" description: List of API keys retrieved. nextPageToken: type: string title: >- Token for fetching the next page of results. Pagination support to be added. TODO: Implement pagination totalSize: type: integer format: int32 description: The total number of API keys. gatewayListAuditLogsResponse: type: object properties: auditLogs: type: array items: type: object $ref: "#/components/schemas/gatewayAuditLogEntry" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of request logs matching the request. gatewayListBatchInferenceJobsResponse: type: object properties: batchInferenceJobs: type: array items: type: object $ref: "#/components/schemas/gatewayBatchInferenceJob" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of batch inference jobs. gatewayListClustersResponse: type: object properties: clusters: type: array items: type: object $ref: "#/components/schemas/gatewayCluster" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of clusters. gatewayListCostsResponse: type: object properties: costDataItems: type: array items: type: object $ref: "#/components/schemas/ListCostsResponseCostDataItem" gatewayListCreditRedemptionsResponse: type: object properties: creditRedemptions: type: array items: type: object $ref: "#/components/schemas/gatewayCreditRedemption" description: The list of credit redemptions. nextPageToken: type: string description: |- A token that can be sent as page_token to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of redemptions matching the request. gatewayListDatasetValidationJobsResponse: type: object properties: datasetValidationJobs: type: array items: type: object $ref: "#/components/schemas/gatewayDatasetValidationJob" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListDatasetsResponse: type: object properties: datasets: type: array items: type: object $ref: "#/components/schemas/gatewayDataset" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of datasets gatewayListDeployedModelsResponse: type: object properties: deployedModels: type: array items: type: object $ref: "#/components/schemas/gatewayDeployedModel" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of deployed models gatewayListDeploymentMetricsResponse: type: object properties: series: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeries" title: Array of time series, each with its own labels and values gatewayListDeploymentShapeVersionsResponse: type: object properties: deploymentShapeVersions: type: array items: type: object $ref: "#/components/schemas/gatewayDeploymentShapeVersion" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of deployment shape versions. gatewayListDeploymentShapesResponse: type: object properties: deploymentShapes: type: array items: type: object $ref: "#/components/schemas/gatewayDeploymentShape" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of deployment shapes. gatewayListDeploymentShardsResponse: type: object properties: deploymentShards: type: array items: type: object $ref: "#/components/schemas/gatewayDeploymentShard" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of deployment shards. gatewayListDeploymentTemplatesResponse: type: object properties: deploymentTemplates: type: array items: type: object $ref: "#/components/schemas/gatewayDeploymentTemplate" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of deployments gatewayListDeploymentsResponse: type: object properties: deployments: type: array items: type: object $ref: "#/components/schemas/gatewayDeployment" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of deployments. gatewayListDeveloperPassesResponse: type: object properties: developerPasses: type: array items: type: object $ref: "#/components/schemas/gatewayDeveloperPass" description: The list of developer passes. nextPageToken: type: string description: Token to retrieve the next page of results. totalSize: type: integer format: int32 description: The total number of developer passes matching the query. gatewayListDpoJobsResponse: type: object properties: dpoJobs: type: array items: type: object $ref: "#/components/schemas/gatewayDpoJob" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of dpo jobs gatewayListEagleTrainingJobsResponse: type: object properties: eagleTrainingJobs: type: array items: type: object $ref: "#/components/schemas/gatewayEagleTrainingJob" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of EAGLE training jobs. gatewayListEvaluationJobsResponse: type: object properties: evaluationJobs: type: array items: type: object $ref: "#/components/schemas/gatewayEvaluationJob" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListEvaluationsResponse: type: object properties: evaluations: type: array items: type: object $ref: "#/components/schemas/gatewayEvaluation" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListEvaluatorVersionsResponse: type: object properties: evaluatorVersions: type: array items: type: object $ref: "#/components/schemas/gatewayEvaluatorVersion" nextPageToken: type: string description: |- A token, which can be sent as page_token to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of evaluator versions. gatewayListEvaluatorsResponse: type: object properties: evaluators: type: array items: type: object $ref: "#/components/schemas/gatewayEvaluator" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListExternalKeysResponse: type: object properties: externalKeys: type: array items: type: object $ref: "#/components/schemas/gatewayExternalKey" description: List of external keys. nextPageToken: type: string description: Next page token. totalSize: type: integer format: int32 description: Total count. gatewayListFeatureFlagsResponse: type: object properties: featureFlags: type: array items: type: object $ref: "#/components/schemas/gatewayFeatureFlag" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListHuggingFaceBillingCostsResponse: type: object properties: requests: type: array items: type: object $ref: "#/components/schemas/gatewayBillingRequestCostRecord" title: Records matching the requested IDs title: Response with billing costs for requested IDs gatewayListIdentityProvidersResponse: type: object properties: identityProviders: type: array items: type: object $ref: "#/components/schemas/gatewayIdentityProvider" title: List of identity providers nextPageToken: type: string title: Next page token totalSize: type: integer format: int32 title: Total count gatewayListInferenceLogsResponse: type: object properties: inferenceLogs: type: array items: type: object $ref: "#/components/schemas/gatewayInferenceLog" title: The list of inference logs nextPageToken: type: string title: A token to retrieve the next page of results totalSize: type: integer format: int32 title: The total number of inference logs matching the request title: ListInferenceLogsResponse returns the list of inference logs gatewayListInvoicesResponse: type: object properties: invoices: type: array items: type: object $ref: "#/components/schemas/gatewayInvoice" title: Usage-based invoices from Orb and Stripe contracts: type: array items: type: object $ref: "#/components/schemas/gatewayBillcomInvoice" title: Contract-based invoices from Bill.com nextPageToken: type: string title: dummy fields to be compatible with list.ListAndPrint totalSize: type: integer format: int32 gatewayListModelsResponse: type: object properties: models: type: array items: type: object $ref: "#/components/schemas/gatewayModel" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of models gatewayListPaymentMethodsResponse: type: object properties: defaultPaymentMethodId: type: string stripePaymentMethods: type: array items: type: object $ref: "#/components/schemas/ListPaymentMethodsResponseStripePaymentMethod" gatewayListPricingPlansResponse: type: object properties: pricingPlans: type: array items: type: object $ref: "#/components/schemas/gatewayPricingPlan" title: The list of pricing plans nextPageToken: type: string title: Token to retrieve the next page of results totalSize: type: integer format: int32 title: The total number of pricing plans matching the query gatewayListQuotasResponse: type: object properties: quotas: type: array items: type: object $ref: "#/components/schemas/gatewayQuota" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of quotas gatewayListReinforcementFineTuningJobsResponse: type: object properties: reinforcementFineTuningJobs: type: array items: type: object $ref: "#/components/schemas/gatewayReinforcementFineTuningJob" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of fine-tuning jobs gatewayListReservationsResponse: type: object properties: reservations: type: array items: type: object $ref: "#/components/schemas/gatewayReservation" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of reservations. gatewayListRlorTrainerJobCheckpointsResponse: type: object properties: checkpoints: type: array items: type: object $ref: "#/components/schemas/gatewayRlorTrainerJobCheckpoint" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListRlorTrainerJobsResponse: type: object properties: rlorTrainerJobs: type: array items: type: object $ref: "#/components/schemas/gatewayRlorTrainerJob" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of fine-tuning jobs gatewayListRoutersResponse: type: object properties: routers: type: array items: type: object $ref: "#/components/schemas/gatewayRouter" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of routers gatewayListSchedulingEntitlementsResponse: type: object properties: schedulingEntitlements: type: array items: type: object $ref: "#/components/schemas/gatewaySchedulingEntitlement" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of entitlements. gatewayListSecretsResponse: type: object properties: secrets: type: array items: type: object $ref: "#/components/schemas/gatewaySecret" nextPageToken: type: string totalSize: type: integer format: int32 description: The total number of secrets. gatewayListServerlessModelsResponse: type: object properties: models: type: array items: type: object $ref: "#/components/schemas/gatewayModel" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of models. description: >- ListServerlessModelsResponse is the serverless catalog read shape: base Models with their serverless_configs nested inside (each config carrying its priced, SPOT-gated serverless_modes). gatewayListSupervisedFineTuningJobsResponse: type: object properties: supervisedFineTuningJobs: type: array items: type: object $ref: "#/components/schemas/gatewaySupervisedFineTuningJob" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of fine-tuning jobs gatewayListTrailsResponse: type: object properties: trails: type: array items: type: object $ref: "#/components/schemas/gatewayTrail" title: The list of trails nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 title: The total number of trails title: Response for listing trails gatewayListTrainingRunsResponse: type: object properties: trainingRuns: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingRun" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListTrainingSessionCheckpointsResponse: type: object properties: trainingSessionCheckpoints: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingSessionCheckpoint" description: The checkpoints in this page, newest first. nextPageToken: type: string description: |- Pass as page_token in a subsequent request to fetch the next page. Empty when there are no more results. totalSize: type: integer format: int32 description: Total number of checkpoints on the session across all pages. gatewayListTrainingSessionChunkOutputsResponse: type: object properties: chunks: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingSessionChunkOutput" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListTrainingSessionTrainingCurvesResponse: type: object properties: curves: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingCurve" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListTrainingSessionsResponse: type: object properties: trainingSessions: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingSession" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListTrainingShapeVersionsResponse: type: object properties: trainingShapeVersions: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingShapeVersion" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListTrainingShapesResponse: type: object properties: trainingShapes: type: array items: type: object $ref: "#/components/schemas/gatewayTrainingShape" nextPageToken: type: string totalSize: type: integer format: int32 gatewayListUserUsageLimitIncreaseRequestsResponse: type: object properties: userUsageLimitIncreaseRequests: type: array items: type: object $ref: "#/components/schemas/gatewayUserUsageLimitIncreaseRequest" nextPageToken: type: string description: >- A token to pass as page_token for the next page. Empty when none remain. totalSize: type: integer format: int32 description: The total number of matching requests. gatewayListUserUsageLimitsResponse: type: object properties: userUsageLimits: type: array items: type: object $ref: "#/components/schemas/gatewayUserUsageLimits" description: >- One entry per user in the account (zero usage for users that incurred none this period), ordered by user id. nextPageToken: type: string description: >- A token to pass as page_token to retrieve the next page. Empty when there are no further pages. gatewayListUsersResponse: type: object properties: users: type: array items: type: object $ref: "#/components/schemas/gatewayUser" nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalSize: type: integer format: int32 description: The total number of users. gatewayLogTrainingSessionMetricsResponse: type: object gatewayMetricPoint: type: object properties: key: type: string description: Metric identifier (e.g. "reward", "loss", "accuracy"). Required. value: type: number format: double description: Observed scalar value at this step. step: type: string format: int64 description: >- Training step the observation was made at. Monotonic per (session, key). timestamp: type: string format: date-time description: >- Optional client-supplied wall-clock timestamp. Defaults to server time. description: >- MetricPoint is a single scalar observation from training. Append-only: the (session, key, step) tuple is the natural key — duplicate writes overwrite the stored value/timestamp so retries from the trainer are safe. gatewayMetrics: type: string enum: - METRICS_UNSPECIFIED - REPLICA_COUNT - AVERAGE_REPLICA_COUNT - LOAD - CONCURRENT_REQUESTS - PROMPT_CACHE_HIT_RATE - REQUESTS_TOTAL - REQUESTS_ERROR_RATE - TOKENS_PROMPT_PER_REQUEST - TOKENS_GENERATED_PER_REQUEST - SPECULATIVE_HIT_TOTAL - SPECULATIVE_HIT_USER - REQUESTS_PER_SECOND - TOKENS_PER_SECOND - TOKENS_GENERATED_PER_SECOND - TOKENS_PROMPT_PER_SECOND - TOKENS_PROMPT_CACHED_PER_SECOND - LATENCY - GENERATION_QUEUE_LATENCY - PREFILL_QUEUE_LATENCY - FIRST_TOKEN_LATENCY - GENERATION_PER_TOKEN_LATENCY - SERVERLESS_REQUESTS_TOTAL - SERVERLESS_REQUESTS_RATE_MIRROR_PER_ACCOUNT - SERVERLESS_REQUESTS_LIMIT_PER_ACCOUNT - SERVERLESS_TOKENS_PROMPT_LIMIT_PER_ACCOUNT - SERVERLESS_TOKENS_PROMPT_RATE_MIRROR_PER_ACCOUNT - SERVERLESS_TOKENS_GENERATED_LIMIT_PER_ACCOUNT - SERVERLESS_TOKENS_GENERATED_RATE_MIRROR_PER_ACCOUNT - SERVERLESS_CACHED_PROMPT_TOKENS - SERVERLESS_TOTAL_PROMPT_TOKENS default: METRICS_UNSPECIFIED description: >- - AVERAGE_REPLICA_COUNT: Time-averaged ready-replica count over the window. - LATENCY: Percentile metrics - SERVERLESS_REQUESTS_TOTAL: Serverless account-specific metrics gatewayModel: type: object properties: name: type: string title: >- The resource name of the model. e.g. accounts/my-account/models/my-model readOnly: true displayName: type: string description: |- Human-readable display name of the model. e.g. "My Model" Must be fewer than 64 characters long. description: type: string description: >- The description of the model. Must be fewer than 1000 characters long. createTime: type: string format: date-time description: The creation time of the model. readOnly: true state: $ref: "#/components/schemas/gatewayModelState" description: The state of the model. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains detailed message when the last model operation fails. readOnly: true kind: $ref: "#/components/schemas/ModelKind" description: |- The kind of model. If not specified, the default is HF_PEFT_ADDON. githubUrl: type: string description: The URL to GitHub repository of the model. huggingFaceUrl: type: string description: The URL to the Hugging Face model. baseModelDetails: $ref: "#/components/schemas/gatewayBaseModelDetails" description: |- Base model details. Required if kind is HF_BASE_MODEL. Must not be set otherwise. peftDetails: $ref: "#/components/schemas/gatewayPEFTDetails" description: |- PEFT addon details. Required if kind is HF_PEFT_ADDON or HF_TEFT_ADDON. teftDetails: $ref: "#/components/schemas/gatewayTEFTDetails" description: |- TEFT addon details. Required if kind is HF_TEFT_ADDON. Must not be set otherwise. public: type: boolean description: If true, the model will be publicly readable. conversationConfig: $ref: "#/components/schemas/gatewayConversationConfig" description: If set, the Chat Completions API will be enabled for this model. contextLength: type: integer format: int32 description: The maximum context length supported by the model. supportsImageInput: type: boolean description: If set, images can be provided as input to the model. supportsTools: type: boolean description: >- If set, tools (i.e. functions) can be provided as input to the model, and the model may respond with one or more tool calls. importedFrom: type: string description: >- The name of the the model from which this was imported. This field is empty if the model was not imported. readOnly: true fineTuningJob: type: string description: >- If the model was created from a fine-tuning job, this is the fine-tuning job name. readOnly: true defaultDraftModel: type: string description: |- The default draft model to use when creating a deployment. If empty, speculative decoding is disabled by default. defaultDraftTokenCount: type: integer format: int32 description: |- The default draft token count to use when creating a deployment. Must be specified if default_draft_model is specified. deployedModelRefs: type: array items: type: object $ref: "#/components/schemas/gatewayDeployedModelRef" description: Populated from GetModel API call only. readOnly: true cluster: type: string description: |- The resource name of the BYOC cluster to which this model belongs. e.g. accounts/my-account/clusters/my-cluster. Empty if it belongs to a Fireworks cluster. readOnly: true deprecationDate: $ref: "#/components/schemas/typeDate" description: >- If specified, this is the date when the serverless deployment of the model will be taken down. calibrated: type: boolean description: >- If true, the model is calibrated and can be deployed to non-FP16 precisions. readOnly: true tunable: type: boolean description: >- Deprecated: V1 training stack only — LoRA only, limited architecture support. If the model has use_training_v2=true, use supervised_lora_tunable and supervised_full_parameter_tunable instead. readOnly: true supportsLora: type: boolean description: Whether this model supports LoRA. useHfApplyChatTemplate: type: boolean description: >- If true, the model will use the Hugging Face apply_chat_template API to apply the chat template. updateTime: type: string format: date-time description: The update time for the model. readOnly: true defaultSamplingParams: type: object additionalProperties: type: number format: float description: >- A json object that contains the default sampling parameters for the model. readOnly: true rlTunable: type: boolean description: >- Deprecated: V1 training stack only — LoRA only, limited architecture support. If the model has use_training_v2=true, use rl_lora_tunable and rl_full_parameter_tunable instead. readOnly: true trainingContextLength: type: integer format: int32 description: The maximum context length supported by the model. snapshotType: $ref: "#/components/schemas/ModelSnapshotType" supportsServerless: type: boolean description: If true, the model has a serverless deployment. readOnly: true useTrainingV2: type: boolean description: >- If true, SFT jobs for this base model use service-mode (StatefulSet + orchestration sidecar) instead of the legacy batch Job path. supervisedLoraTunable: type: boolean description: >- V2 only. Whether the model supports LoRA supervised fine-tuning and DPO (lora_rank > 0). True when a validated LORA_TRAINER training shape exists. readOnly: true supervisedFullParameterTunable: type: boolean description: >- V2 only. Whether the model supports full-parameter supervised fine-tuning and DPO (lora_rank = 0). True when a validated POLICY_TRAINER training shape exists. readOnly: true rlLoraTunable: type: boolean description: >- V2 only. Whether the model supports LoRA reinforcement learning (lora_rank > 0). True when a validated LORA_TRAINER training shape exists plus a deployment shape. readOnly: true rlFullParameterTunable: type: boolean description: >- V2 only. Whether the model supports full-parameter reinforcement learning (lora_rank = 0). True when validated POLICY_TRAINER + LORA_TRAINER training shapes exist plus a deployment shape. readOnly: true encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true serverlessModes: type: array items: type: object $ref: "#/components/schemas/gatewayServerlessMode" description: >- The serverless modes this base model is available on — the serverless read shape. Populated on read (e.g. the serverless catalog read path); one entry per serving mode. Empty for non-serverless models. readOnly: true gatewayModelState: type: string enum: - STATE_UNSPECIFIED - UPLOADING - READY default: STATE_UNSPECIFIED description: |- - UPLOADING: The model is still being uploaded (upload is asynchronous). - READY: The model is ready to be used. gatewayModelVersion: type: object properties: name: type: string description: The resource name of the deployment snapshot. readOnly: true createTime: type: string format: date-time readOnly: true snapshot: $ref: "#/components/schemas/gatewayModel" readOnly: true gatewayMultiRegion: type: string enum: - MULTI_REGION_UNSPECIFIED - GLOBAL - US - EUROPE - APAC default: MULTI_REGION_UNSPECIFIED gatewayNotificationSettings: type: object properties: monthlySpendThresholds: type: array items: type: object $ref: "#/components/schemas/typeMoney" description: >- Spend thresholds at which to send monthly usage warning emails. These should be below the account's monthly spend quota. Example: [{currency_code: "USD", units: 500}, {currency_code: "USD", units: 800}] would alert at $500 and $800. Note: An alert at 80% of the monthly spend limit is always added in addition to the thresholds specified in this list. description: >- NotificationSettings configures notification preferences for an account. (accounts/{account}/notificationSettings) when more notification categories are added. Requires adding singleton support to the resource generator. gatewayOidcConfig: type: object properties: issuerUrl: type: string title: OIDC issuer URL clientId: type: string title: Client ID clientSecret: type: string title: Client secret title: OIDC Configuration required: - issuerUrl - clientId - clientSecret gatewayOrpoConfig: type: object properties: lambda: type: number format: float description: Weight for the ORPO odds-ratio loss term. description: Hyperparameters for Odds Ratio Preference Optimization (ORPO) training. gatewayPEFTDetails: type: object properties: baseModel: type: string title: The base model name. e.g. accounts/fireworks/models/falcon-7b r: type: integer format: int32 description: |- The rank of the update matrices. Must be between 4 and 64, inclusive. targetModules: type: array items: type: string title: >- This is the target modules for an adapter that we extract from for more information what target module means, check out https://huggingface.co/docs/peft/conceptual_guides/lora#common-lora-parameters-in-peft baseModelType: type: string description: The type of the model. readOnly: true mergeAddonModelName: type: string title: >- The resource name of the model to merge with base model, e.g accounts/fireworks/models/falcon-7b-lora title: PEFT addon details. required: - baseModel - r - targetModules gatewayPlacement: type: object properties: region: $ref: "#/components/schemas/gatewayRegion" description: The region where the deployment must be placed. multiRegion: $ref: "#/components/schemas/gatewayMultiRegion" description: The multi-region where the deployment must be placed. regions: type: array items: $ref: "#/components/schemas/gatewayRegion" title: The list of regions where the deployment must be placed description: >- The desired geographic region where the deployment must be placed. Exactly one field will be specified. gatewayPolicySettings: type: object properties: name: type: string title: Resource name, e.g. accounts/my-account/policySettings readOnly: true defaultPermissions: $ref: "#/components/schemas/PolicySettingsModelPermissions" description: |- Per-account default permissions. If unset, defaults to allow-all. This powers the UI's "Default" row. rules: type: array items: type: object $ref: "#/components/schemas/PolicySettingsModelAccessRule" description: >- Per-model override rows. A model not listed here uses default_permissions. updateTime: type: string format: date-time readOnly: true cmekRequired: type: boolean description: >- Whether new customer-data resources for this account must be CMEK-encrypted. Readable by the account, but only a Fireworks superuser can change it (the write path gates this explicitly). Enabling requires a READY primary ExternalKey. description: >- Account-level policy settings (singleton per account). Holds model access and may grow with other policy sections (e.g. regional residency) without separate top-level API resources. gatewayPreviewDatasetResponse: type: object properties: examples: type: array items: type: object $ref: "#/components/schemas/gatewayExample" description: The list of examples in the dataset for the requested page. nextPageToken: type: string description: >- A token, which can be sent as `page_token` to retrieve the next page. If this field is omitted, there are no subsequent pages. totalCount: type: integer format: int32 description: The total number of examples in the dataset. gatewayPreviewEvaluationResponse: type: object properties: results: type: array items: type: object $ref: "#/components/schemas/gatewayPreviewEvaluationResult" title: Results for each sample totalSamples: type: integer format: int32 title: Summary statistics totalRuntimeMs: type: string format: int64 title: overalll runtime title: Response for preview evaluation gatewayPreviewEvaluationResult: type: object properties: success: type: boolean title: The input sample reason: type: string title: Output from the evaluation run score: type: number format: double description: |- Score (if applicable) Deprecated: Use metrics field instead. messages: type: array items: type: object title: messages, which can be any kind of object metrics: type: object additionalProperties: type: number format: double title: Metrics from the evaluation run title: A single evaluation result from a preview gatewayPreviewEvaluatorResponse: type: object properties: results: type: array items: type: object $ref: "#/components/schemas/gatewayPreviewEvaluatorSampleResult" totalSamples: type: integer format: int32 totalRuntimeMs: type: string format: int64 stdout: type: array items: type: string stderr: type: array items: type: string gatewayPreviewEvaluatorSampleResult: type: object properties: success: type: string title: >- Overall eval success (true) or failure (false) on the sample, empty string means success not defined score: type: number format: double title: Score (rollup score if applicable) perMetricEvals: type: object additionalProperties: type: object title: Per metric eval results reason: type: string title: reason for the eval result gatewayPricingPlan: type: object properties: name: type: string title: |- The resource name of the pricing plan Format: accounts/{account}/pricingPlans/{pricing_plan_id} readOnly: true billingType: $ref: "#/components/schemas/PricingPlanBillingType" title: Billing type determines how usage is tracked and billed tokenConfig: type: array items: type: object $ref: "#/components/schemas/PricingPlanTokenBillingConfig" title: Configuration for token-based billing acceleratorHourConfig: type: array items: type: object $ref: "#/components/schemas/PricingPlanAcceleratorHourBillingConfig" title: Configuration for accelerator-hour billing startTime: type: string format: date-time title: When this pricing plan becomes active endTime: type: string format: date-time title: When this pricing plan expires createTime: type: string format: date-time title: Creation timestamp readOnly: true updateTime: type: string format: date-time title: Last update timestamp readOnly: true state: $ref: "#/components/schemas/gatewayPricingPlanState" title: The state of the pricing plan readOnly: true description: >- PricingPlan defines a custom billing plan for an account Each PricingPlan is associated with a contract, and describes the pricing terms for a specific service. required: - billingType - startTime gatewayPricingPlanState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY default: STATE_UNSPECIFIED gatewayPromoteCheckpointResponse: type: object properties: model: $ref: "#/components/schemas/gatewayModel" description: The promoted model. Present only for synchronous promotion. operation: $ref: "#/components/schemas/googlelongrunningOperation" description: >- Long-running promotion operation. Present only for async promotion. Poll GetOperation on operation.name until done; the promoted model is returned in the operation response. gatewayPromoteTrainingSessionCheckpointResponse: type: object properties: model: $ref: "#/components/schemas/gatewayModel" gatewayProvider: type: object properties: id: type: string title: Provider ID (e.g., "openai:gpt-4") config: type: object additionalProperties: type: string title: Optional provider-specific configuration label: type: string title: Optional label for the provider title: Represents an LLM provider gatewayPurpose: type: string enum: - PURPOSE_UNSPECIFIED - PURPOSE_PILOT default: PURPOSE_UNSPECIFIED description: Scheduling purpose for training jobs and deployments. gatewayQuota: type: object properties: name: type: string title: >- The resource name of the quota, e.g. accounts/my-account/quotas/h100-us-iowa-1 readOnly: true value: type: string format: int64 description: >- The value of the quota being enforced. This may be lower than the max_value if the user manually lowers it. maxValue: type: string format: int64 description: The maximum approved value. usage: type: number format: double description: The usage of the quota. readOnly: true updateTime: type: string format: date-time description: The update time for the quota. readOnly: true gatewayRedeemCreditCodeRequest: type: object properties: code: type: string description: The user-facing credit code string (e.g., "HACKATHON_2025"). name: type: string description: The resource name of the account redeeming the credit code. description: >- RedeemCreditCodeRequest is the request to redeem a credit code for an account. required: - code - name gatewayRedeemCreditCodeResponse: type: object properties: amount: $ref: "#/components/schemas/typeMoney" description: The amount of credits redeemed. description: RedeemCreditCodeResponse returns the amount of credits redeemed. gatewayRefreshSessionTokenRequest: type: object properties: refreshToken: type: string description: The refresh token used to mint a fresh access token. description: Request to refresh an existing session JWT before it expires. gatewayRefreshSessionTokenResponse: type: object properties: token: type: string description: New signed JWT with fresh TTL and the same session claims. expireTime: type: string format: date-time description: When the new token expires. refreshToken: type: string description: |- Refresh token for subsequent refreshes. In V1 this may be unchanged; later versions may rotate it. refreshExpireTime: type: string format: date-time description: When the refresh token expires. description: Response containing the refreshed session JWT. gatewayRegion: type: string enum: - REGION_UNSPECIFIED - US_IOWA_1 - US_VIRGINIA_1 - US_VIRGINIA_2 - US_ILLINOIS_1 - AP_TOKYO_1 - US_ARIZONA_1 - US_TEXAS_1 - US_ILLINOIS_2 - EU_FRANKFURT_1 - US_TEXAS_2 - EU_ICELAND_1 - EU_ICELAND_2 - US_WASHINGTON_1 - US_WASHINGTON_2 - US_WASHINGTON_3 - AP_TOKYO_2 - US_CALIFORNIA_1 - US_UTAH_1 - US_ARIZONA_3 - US_GEORGIA_1 - US_GEORGIA_2 - US_WASHINGTON_4 - US_GEORGIA_3 - NA_BRITISHCOLUMBIA_1 - US_GEORGIA_4 - US_OHIO_1 - US_NEWYORK_1 - EU_NETHERLANDS_1 - US_WASHINGTON_5 - US_MINNESOTA_1 - US_CALIFORNIA_2 - NA_BRITISHCOLUMBIA_2 - AP_MALAYSIA_2 - US_OREGON_1 - NA_BRITISHCOLUMBIA_3 - AP_NEWSOUTHWALES_1 default: REGION_UNSPECIFIED gatewayReinforcementFineTuningJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true completedTime: type: string format: date-time description: The completed time for the reinforcement fine-tuning job. readOnly: true dataset: type: string description: The name of the dataset used for training. evaluationDataset: type: string description: The name of a separate dataset to use for evaluation. evalAutoCarveout: type: boolean description: Whether to auto-carve the dataset for eval. state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true createdBy: type: string description: The email address of the user who initiated this fine-tuning job. readOnly: true trainingConfig: $ref: "#/components/schemas/gatewayBaseTrainingConfig" description: Common training configurations. evaluator: type: string description: The evaluator resource name to use for RLOR fine-tuning job. wandbConfig: $ref: "#/components/schemas/gatewayWandbConfig" description: >- The Weights & Biases team/user account for logging training progress. awsS3Config: $ref: "#/components/schemas/gatewayAwsS3Config" description: The AWS configuration for S3 dataset access. azureBlobStorageConfig: $ref: "#/components/schemas/gatewayAzureBlobStorageConfig" description: The Azure configuration for Blob Storage dataset access. outputStats: type: string description: The output dataset's aggregated stats for the evaluation job. readOnly: true jobProgress: $ref: "#/components/schemas/gatewayJobProgress" description: Job progress. readOnly: true inferenceParameters: $ref: >- #/components/schemas/gatewayReinforcementFineTuningJobInferenceParameters description: RFT inference parameters. chunkSize: type: integer format: int32 description: >- Data chunking for rollout, default size 200, enabled when dataset > 300. Valid range is 1-10,000. outputMetrics: type: string readOnly: true maxInferenceReplicaCount: type: integer format: int32 title: |- Maximum number of replicas to use for the deployment. Default is 1 nodeCount: type: integer format: int32 description: |- The number of nodes to use for the fine-tuning job. If not specified, the default is 1. lossConfig: $ref: "#/components/schemas/gatewayReinforcementLearningLossConfig" description: >- Reinforcement learning loss method + hyperparameters for the underlying trainers. trainerLogsSignedUrl: type: string description: |- The signed URL for the trainer logs file (stdout/stderr). Only populated if the account has trainer log reading enabled. readOnly: true acceleratorSeconds: type: object additionalProperties: type: string format: int64 description: >- Accelerator seconds used by the job, keyed by accelerator type (e.g., "NVIDIA_H100_80GB"). Updated when job completes or is cancelled. readOnly: true maxConcurrentRollouts: type: integer format: int32 description: Maximum number of concurrent rollouts during the RFT job. maxConcurrentEvaluations: type: integer format: int32 description: Maximum number of concurrent evaluations during the RFT job. purpose: $ref: "#/components/schemas/gatewayPurpose" description: Scheduling purpose for this job. useReservation: type: boolean description: >- Whether to try the account's reservation capacity before falling back to shared trainer capacity. encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true required: - dataset - evaluator gatewayReinforcementFineTuningJobInferenceParameters: type: object properties: maxOutputTokens: type: integer format: int32 description: Maximum number of tokens to generate per response. temperature: type: number format: float description: Sampling temperature, typically between 0 and 2. topP: type: number format: float description: Top-p sampling parameter, typically between 0 and 1. responseCandidatesCount: type: integer format: int32 title: >- Number of response candidates to generate per input. RFT requires at least 2 candidates extraBody: type: string description: |- Additional parameters for the inference request as a JSON string. For example: "{\"stop\": [\"\\n\"]}". topK: type: integer format: int32 description: >- Top-k sampling parameter, limits the token selection to the top k tokens. title: RFT inference parameters gatewayReinforcementLearningLossConfig: type: object properties: method: $ref: "#/components/schemas/ReinforcementLearningLossConfigMethod" klBeta: type: number format: float description: |- KL coefficient (beta) override for GRPO-like methods. If unset, the trainer default is used. dpo: $ref: "#/components/schemas/gatewayDpoConfig" description: DPO-specific configuration. Intended for METHOD=DPO. orpo: $ref: "#/components/schemas/gatewayOrpoConfig" description: ORPO-specific configuration. Intended for METHOD=ORPO. description: >- Loss method + hyperparameters for reinforcement-learning-style fine-tuning (e.g. RFT / RL trainers). For preference jobs (DPO API), the default loss method is GRPO when METHOD_UNSPECIFIED. gatewayReplicaStats: type: object properties: pendingSchedulingReplicaCount: type: integer format: int32 description: Number of replicas waiting to be scheduled to a node. readOnly: true downloadingModelReplicaCount: type: integer format: int32 description: Number of replicas downloading model weights. readOnly: true initializingReplicaCount: type: integer format: int32 description: Number of replicas initializing the model server. readOnly: true readyReplicaCount: type: integer format: int32 description: Number of replicas that are ready and serving traffic. readOnly: true revocableReplicaCount: type: integer format: int32 title: The number of replicas that can be revoked readOnly: true effectiveReplicaCount: type: number format: float description: >- The effective number of replicas currently serving traffic, including fractional progress toward a replica that is not yet fully ready. Equals ready_replica_count when no replica is partially ready. readOnly: true gatewayReservation: type: object properties: name: type: string title: >- The resource name of the reservation. e.g. accounts/my-account/reservations/abcdef readOnly: true displayName: type: string description: >- Human-readable display name of the reservation. e.g. "My Reservation" Must be fewer than 64 characters long. description: type: string description: Description of the reservation. createTime: type: string format: date-time description: The creation time of the reservation. readOnly: true createdBy: type: string description: The email address of the user who created this reservation. readOnly: true region: $ref: "#/components/schemas/gatewayRegion" description: >- The region for this reservation. Exactly one of region or multi_region must be specified. multiRegion: $ref: "#/components/schemas/gatewayMultiRegion" description: |- The multi-region for this reservation. Exactly one of region or multi_region must be specified. acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: The type of accelerator for this reservation. reservedCount: type: integer format: int32 description: The number of accelerators reserved. startTime: type: string format: date-time description: >- The timestamp at which the reservation starts (inclusive). If unspecified, this is midnight of the following calendar day (UTC). endTime: type: string format: date-time description: |- The timestamp at which the reservation ends (exclusive). If unspecified, this will be 1 calendar year after the start time. updateTime: type: string format: date-time description: The update time for the reservation. readOnly: true state: $ref: "#/components/schemas/gatewayReservationState" description: The lifecycle state of the reservation. readOnly: true required: - acceleratorType - reservedCount gatewayReservationState: type: string enum: - STATE_UNSPECIFIED - ACTIVE - EXPIRED - DELETING - DELETED - UPDATING default: STATE_UNSPECIFIED description: >- The lifecycle state of the reservation. Live states (UPDATING/ACTIVE) reflect fulfillment (whether every shard's reserved_count is materialized on nodes); terminal states are driven by expiry or delete: expiry (time-driven): while live past end_time the watcher releases the reserved nodes, then commits EXPIRED once released. delete (admin-driven): DELETING is a stored intent while the watcher releases nodes, then DELETED once released. The authoritative expiry check is always `end_time < now`, never this value. STATE_UNSPECIFIED on pre-existing rows is treated as ACTIVE. - ACTIVE: Live and fulfilled: every shard has its reserved_count of GPUs fulfilled on nodes. Within [start_time, end_time) it grants capacity. - EXPIRED: end_time has passed and all reserved nodes are released (terminal). The row is kept permanently as a contract record. - DELETING: Deletion was requested; will release the reserved nodes. - DELETED: Deletion is complete: no nodes carry this reservation's capacity. - UPDATING: Live but not yet fulfilled: the reservation was just created or rematerialized, or its pool is short, so at least one shard does not yet have its reserved_count fulfilled. The watcher commits ACTIVE once every shard is filled. gatewayRlorTrainerJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true completedTime: type: string format: date-time readOnly: true dataset: type: string description: The name of the dataset used for training. evaluationDataset: type: string description: The name of a separate dataset to use for evaluation. evalAutoCarveout: type: boolean description: Whether to auto-carve the dataset for eval. state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true createdBy: type: string description: The email address of the user who initiated this fine-tuning job. readOnly: true trainingConfig: $ref: "#/components/schemas/gatewayBaseTrainingConfig" description: Common training configurations. rewardWeights: type: array items: type: string description: >- A list of reward metrics to use for training in format of "=". wandbConfig: $ref: "#/components/schemas/gatewayWandbConfig" description: >- The Weights & Biases team/user account for logging training progress. awsS3Config: $ref: "#/components/schemas/gatewayAwsS3Config" description: The AWS configuration for S3 dataset access. azureBlobStorageConfig: $ref: "#/components/schemas/gatewayAzureBlobStorageConfig" description: The Azure configuration for Azure Blob Storage dataset access. jobProgress: $ref: "#/components/schemas/gatewayJobProgress" description: Job progress. readOnly: true keepAlive: type: boolean title: indicates this RLOR trainer job should run in keep-alive mode rolloutDeploymentName: type: string description: >- Rollout deployment name associated with this RLOR trainer job. This is optional. If not set, trainer will not trigger weight sync to rollout engine. lossConfig: $ref: "#/components/schemas/gatewayReinforcementLearningLossConfig" description: >- Reinforcement learning loss method + hyperparameters for the underlying trainer. nodeCount: type: integer format: int32 description: |- The number of nodes to use for the fine-tuning job. If not specified, the default is 1. acceleratorSeconds: type: object additionalProperties: type: string format: int64 description: >- Accelerator seconds used by the job, keyed by accelerator type (e.g., "NVIDIA_H100_80GB"). Updated periodically. readOnly: true serviceMode: type: boolean title: >- Whether to deploy as a service with tinker-style api endpoints exposure directRouteHandle: type: string title: |- Only valid when service_mode enabled The direct route handle for the trainer in service mode (tinker api) readOnly: true hotLoadDeploymentId: type: string description: >- The deployment ID used for hot loading. When set, checkpoints are saved to this deployment's hot load bucket, enabling weight swaps on inference. Only valid for service-mode or keep-alive jobs. purpose: $ref: "#/components/schemas/gatewayPurpose" description: Scheduling purpose for this job. forwardOnly: type: boolean description: >- When true, run the trainer in forward-only mode (no backward/optimizer). Used for reference models in GRPO that only need forward passes. managedBy: type: string description: For managed service use only. Users do not need to set this field. trainerReplicaCount: type: integer format: int32 description: >- Data-parallel replica count for service-mode trainers. When unset or 0, defaults to 1. Values greater than 1 launch replicated HSDP training. inactivityTimeout: type: string description: >- Trainer inactivity timeout. The trainer reports tracked activity, including trainer API operations and active-session heartbeats. If no tracked activity is observed for this duration, the trainer is automatically stopped. When unset or 0, defaults to 60 minutes. Set disable_inactivity_cleanup to true to disable automatic cleanup. GPU usage continues to accrue while the trainer is running. disableInactivityCleanup: type: boolean description: >- Disable trainer inactivity cleanup. When true, the trainer is not automatically stopped due to inactivity. GPU usage continues to accrue while the trainer is running. encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true gatewayRlorTrainerJobCheckpoint: type: object properties: name: type: string title: >- The resource name of the checkpoint. Format: accounts/{account}/rlorTrainerJobs/{job}/checkpoints/{checkpoint} readOnly: true createTime: type: string format: date-time description: The creation time of the checkpoint (from GCS object metadata). readOnly: true updateTime: type: string format: date-time description: |- The update time of the checkpoint. Equal to create_time since checkpoints are immutable GCS objects. readOnly: true checkpointType: $ref: "#/components/schemas/gatewayCheckpointType" description: The type of checkpoint. readOnly: true promotable: type: boolean description: >- Whether this checkpoint can be promoted to a model. Only INFERENCE_BASE and INFERENCE_LORA checkpoints are promotable. INFERENCE_ARC_V2 and TRAINING_* checkpoints cannot be promoted directly. readOnly: true restoreReference: type: string title: |- Opaque checkpoint reference accepted by training restore APIs. Example: cross_job://{RlorTrainerJobId}/{CheckpointId} readOnly: true availability: type: array items: type: object $ref: "#/components/schemas/gatewayCheckpointAvailability" description: >- Storage availability known for this checkpoint. Physical storage URIs are intentionally not exposed; use restore_reference for restore workflows. readOnly: true description: >- A checkpoint saved during an RLOR trainer job. Checkpoints are stored in GCS and discovered on-demand (not persisted in the database). gatewayRolloutStrategy: type: string enum: - ROLLOUT_STRATEGY_UNSPECIFIED - ROLLOUT_STRATEGY_STANDARD - ROLLOUT_STRATEGY_HOT_RELOAD default: ROLLOUT_STRATEGY_UNSPECIFIED description: |- The rollout strategy to use when deploying the model version. - ROLLOUT_STRATEGY_STANDARD: Standard rollout strategy updates the deployment using a normal k8s rolling restart - ROLLOUT_STRATEGY_HOT_RELOAD: Hot reload rollout strategy updates the deployment by hot reloading the model version on the existing replicas of the deployment gatewayRouter: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true state: $ref: "#/components/schemas/gatewayRouterState" description: The state of the router. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true deployments: type: array items: type: string description: The deployment names to be covered by the router. model: type: string description: |- The model name to route requests to. model is only applicable to single-region deployments. For multi-region deployments, model must be empty. weightedRandom: $ref: "#/components/schemas/gatewayWeightedRandomStrategy" evenLoad: $ref: "#/components/schemas/gatewayEvenLoadStrategy" aliases: type: array items: type: string title: >- Aliases for the router. These are the alias names that can be used to route requests through the router. By default, these aliases are generated by the system. Examples: "accounts/foo/deployments/my-deployment" or "foo/my-deployment" "accounts/foo/routers/my-router" "accounts/foo/models/my-base-model" - only if this is the first deployment using this model readOnly: true autoGenerated: type: boolean description: >- auto_generated indicates this router was generated by the system for a deployment. Auto-generated routers cannot be updated or deleted by the user. readOnly: true public: type: boolean description: >- True if the router is public (any account can query the underlying workload), false if the router is private (only the account that owns the router can query the underlying workload). gatewayRouterState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - UPDATING - DELETING default: STATE_UNSPECIFIED description: |- - CREATING: The router is being created. - READY: The router is ready for access. - UPDATING: There are in-progress updates happening with the router. - DELETING: The router is being deleted. gatewaySKUInfo: type: object properties: sku: type: string amount: $ref: "#/components/schemas/typeMoney" unit: type: string title: |- e.g. 1M tokens, hour gatewaySamlConfig: type: object properties: metadataUrl: type: string title: SAML metadata URL (must be publicly accessible) metadataXml: type: string title: >- SAML metadata XML content (use when metadata URL is not publicly accessible) samlEmailAttribute: type: string description: >- The SAML attribute name that contains the user's email address. Different identity providers use different attribute names for email: - Azure AD / ADFS / Okta: "http://schemas.xmlsoap.org/ws/2005/05/identity/claims/emailaddress" (default) - Rippling / some SAML IdPs: "email" - X.500 / LDAP-based IdPs: "urn:oid:1.2.840.113549.1.9.1" If not set, defaults to "http://schemas.xmlsoap.org/ws/2005/05/identity/claims/emailaddress". description: |- SAML Configuration Exactly one of metadata_url or metadata_xml must be provided. gatewayScalingSchedule: type: object properties: minReplicaCount: type: integer format: int32 description: >- Minimum number of replicas guaranteed when this schedule is active. When multiple schedules overlap, the effective minimum is the highest min_replica_count across all active schedules ("max wins"). Must be >= 0 and <= the deployment's max_replica_count. schedule: type: string description: >- Cron expression defining when this schedule's window starts. Standard 5-field cron format: minute hour day-of-month month day-of-week. Examples: "0 8 * * Mon-Fri" (8am weekdays), "0 0 1 * *" (midnight on 1st of month). duration: type: string description: >- Duration that the schedule remains active after the cron trigger. Must be between 300 seconds (5 minutes) and 604,800 seconds (7 days), and use whole-second precision. Schedules needing longer windows should raise the deployment's base min_replica_count instead. Example: "36000s" = 10 hours (e.g., 8am to 6pm). timezone: type: string description: >- IANA timezone for the cron expression. e.g., "America/New_York", "Europe/London", "UTC". Required because cron expressions without a timezone are ambiguous. DST transitions are handled automatically. description: type: string description: |- Human-readable description of the schedule. e.g., "Weekday business hours", "Wednesday peak load". disabled: type: boolean description: >- If true, this schedule is temporarily disabled without being deleted. Useful for holidays or temporary schedule changes. title: >- A time-based scaling schedule that sets a minimum replica floor during a recurring window. Follows the GCP MIG scaling schedule pattern: cron start time + duration. required: - minReplicaCount - schedule - duration - timezone gatewaySchedulingEntitlement: type: object properties: name: type: string title: >- The resource name. e.g. accounts/my-account/schedulingEntitlements/abcdef readOnly: true acceleratorCount: type: integer format: int32 description: >- Maximum number of accelerators the account may draw concurrently from the pool. acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: Accelerator type (SKU) this entitlement is scoped to. timeToFulfill: type: string description: >- Provisioning SLA (H): Fireworks schedules up to accelerator_count accelerators within this duration of a valid request. Must be at least 48h for now — a longer SLA is less restrictive. timeout: type: string description: >- Guaranteed uninterrupted run time (H2): once scheduled, a job runs at least this long. After it, the job becomes evictable under load (a held evict-to-admit right, exercised only under contention); jobs may run longer when there is no contention. Must be exactly 48h for now — the only value currently supported. state: $ref: "#/components/schemas/gatewaySchedulingEntitlementState" readOnly: true createTime: type: string format: date-time description: The creation time of the schedulingEntitlement. readOnly: true updateTime: type: string format: date-time description: The last update time of the schedulingEntitlement. readOnly: true createdBy: type: string description: >- The email address of the user who created this schedulingEntitlement. readOnly: true description: >- SchedulingEntitlement is a per-account, pay-per-usage record granting an account the right to draw up to `accelerator_count` accelerators of `accelerator_type` from a single fungible training pool, with jobs scheduled within `time_to_fulfill` (H) and guaranteed to run at least `timeout` (H2). It reserves no capacity: nothing is materialized or held. The scheduler reads these fields; this resource only stores them. The governing contract is stored externally; `contract` links to it. required: - acceleratorCount - acceleratorType - timeToFulfill - timeout gatewaySchedulingEntitlementState: type: string enum: - STATE_UNSPECIFIED - ACTIVE - CANCELLED default: STATE_UNSPECIFIED description: >- The lifecycle state of the schedulingEntitlement. There is no time-based expiry. - STATE_UNSPECIFIED: Default zero value; never a stored state. - ACTIVE: Live: the schedulingEntitlement grants capacity and jobs may draw against it. - CANCELLED: Terminated via delete; the row is retained as a contract record. gatewaySecret: type: object properties: name: type: string title: |- name follows the convention accounts/account-id/secrets/unkey-key-id keyName: type: string title: name of the key. In this case, it can be WOLFRAM_ALPHA_API_KEY value: type: string example: sk-1234567890abcdef description: >- The secret value. This field is INPUT_ONLY and will not be returned in GET or LIST responses for security reasons. The value is only accepted when creating or updating secrets. required: - name - keyName gatewayServerlessMode: type: object properties: name: type: string description: >- The resource name, e.g. accounts/fireworks/models/kimi-k2p6/serverlessModes/fast. {AccountId} must be a serverless account (fireworks, ...); {ModelId} is the base model's id; {ServerlessModeId} is the mode name (default | priority | fast | spot). readOnly: true usageIdentifier: type: string description: >- Invocation recipe: the value to pass as "model" in inference requests. Empty means use the base model id. May reference a DIFFERENT resource than the base model — a model OR a router, e.g. accounts/fireworks/routers/kimi-k2p6-fast for fast. serviceTier: type: string description: >- Invocation recipe: the service_tier request-body flag layered on top of the model id (e.g. "priority"). Only the priority mode sets this today. skuInfos: type: array items: type: object $ref: "#/components/schemas/gatewaySKUInfo" description: >- Per-path token pricing, sourced live from Orb on the read path. Reuses the same SKUInfo type as Model.sku_infos (e.g. "LLM input tokens (uncached)", "LLM input tokens (cached)", "LLM output tokens"; google.type.Money at unit="1M tokens"). Resolved per tier (DEFAULT/PRIORITY/FAST/...). Empty on the stored config; only populated by ListServerlessModels. (api_only) keeps this derived field out of the database entirely (like Model.sku_infos): it is never persisted by the admin write path and never read back from storage, only computed live at read time. readOnly: true useCases: type: array items: type: string description: Curated use-case tags driving discovery filters (e.g. "coding"). updatedBy: type: string description: The user who last mutated this mode (audit). readOnly: true createTime: type: string format: date-time readOnly: true updateTime: type: string format: date-time readOnly: true description: >- ServerlessMode is one way a serverless base model can be invoked — standard serverless (default), Priority, Fast, or Spot — together with the invocation recipe customers must use for it. It is a child resource of the base Model: the {ServerlessModeId} segment of the resource name IS the mode name (default | priority | fast | spot). The base model is the linking key for all of its serving modes, even when a mode is served through a different resource (a Fast router, a Spot deployment). Modes split into two classes: - Alternate-resource modes (usage_identifier set) — e.g. fast resolves to a router, spot to the spot deployment's own model id. - Flag modes (service_tier set) — e.g. priority uses the base model id plus a request-body service_tier flag. Pricing is sourced live from Orb (keyed by the resolved model name / tier), so no price is stored on the mode. gatewaySplitDatasetResponse: type: object properties: chunkDatasetNames: type: array items: type: string title: The resource names of the created chunk datasets chunksCreated: type: integer format: int32 title: The number of chunks created totalExamples: type: string format: int64 title: The total number of examples processed title: Response message for dataset splitting gatewaySplitted: type: object properties: sourceDatasetId: type: string required: - sourceDatasetId gatewayStatus: type: object properties: code: $ref: "#/components/schemas/gatewayCode" description: The status code. message: type: string description: A developer-facing error message in English. title: >- Mimics [https://github.com/googleapis/googleapis/blob/master/google/rpc/status.proto] gatewaySupervisedFineTuningJob: type: object properties: name: type: string readOnly: true displayName: type: string createTime: type: string format: date-time readOnly: true completedTime: type: string format: date-time readOnly: true dataset: type: string description: The name of the dataset used for training. awsS3Config: $ref: "#/components/schemas/gatewayAwsS3Config" description: The AWS configuration for S3 dataset access. azureBlobStorageConfig: $ref: "#/components/schemas/gatewayAzureBlobStorageConfig" description: The Azure configuration for Azure Blob Storage dataset access. state: $ref: "#/components/schemas/gatewayJobState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true createdBy: type: string description: The email address of the user who initiated this fine-tuning job. readOnly: true outputModel: type: string description: >- The model ID to be assigned to the resulting fine-tuned model. If not specified, the job ID will be used. baseModel: type: string description: |- The name of the base model to be fine-tuned Only one of 'base_model' or 'warm_start_from' should be specified. warmStartFrom: type: string description: |- The PEFT addon model in Fireworks format to be fine-tuned from Only one of 'base_model' or 'warm_start_from' should be specified. jinjaTemplate: type: string description: >- Deprecated: literal Jinja templates are not supported by Training V2. Conversation rendering is selected from the base model's registered renderer configuration instead. earlyStop: type: boolean description: "Deprecated: early stopping is not supported by managed training." epochs: type: integer format: int32 description: The number of epochs to train for. learningRate: type: number format: float description: The learning rate used for training. maxContextLength: type: integer format: int32 description: The maximum context length to use with the model. loraRank: type: integer format: int32 description: The rank of the LoRA layers. wandbConfig: $ref: "#/components/schemas/gatewayWandbConfig" description: >- The Weights & Biases team/user account for logging training progress. evaluationDataset: type: string description: The name of a separate dataset to use for evaluation. isTurbo: type: boolean description: Whether to run the fine-tuning job in turbo mode. evalAutoCarveout: type: boolean description: Whether to auto-carve the dataset for eval. updateTime: type: string format: date-time description: The update time for the supervised fine-tuning job. readOnly: true nodes: type: integer format: int32 description: |- Deprecated: multi-node scheduling is now handled by the cookbook orchestrator in V2 workflows. This field is ignored for V2 jobs and will be removed in a future release. batchSize: type: integer format: int32 description: >- Deprecated: legacy V1 token budget. Training V2 batches by samples via batch_size_samples. mtpEnabled: type: boolean description: >- Deprecated: MTP is no longer supported by managed training. This field is retained for API wire compatibility only. mtpNumDraftTokens: type: integer format: int32 description: "Deprecated: see mtp_enabled." mtpFreezeBaseModel: type: boolean description: "Deprecated: see mtp_enabled." jobProgress: $ref: "#/components/schemas/gatewayJobProgress" description: Job progress. readOnly: true metricsFileSignedUrl: type: string title: The signed URL for the metrics file trainerLogsSignedUrl: type: string description: |- The signed URL for the trainer logs file (stdout/stderr). Only populated if the account has trainer log reading enabled. readOnly: true renderSamplesSignedUrl: type: string description: The signed URL for orchestrator-rendered token IDs and loss masks. readOnly: true gradientAccumulationSteps: type: integer format: int32 description: |- Deprecated: legacy V1 gradient accumulation. Training V2 batches by samples via batch_size_samples and rejects this field when set. learningRateWarmupSteps: type: integer format: int32 title: Number of steps for learning rate warm up lrScheduler: $ref: "#/components/schemas/gatewayLearningRateScheduler" description: >- The learning-rate schedule (constant/linear/cosine + per-type knobs). When unset, the trainer uses the legacy constant schedule. batchSizeSamples: type: integer format: int32 description: The number of samples per gradient batch. estimatedCost: $ref: "#/components/schemas/typeMoney" description: The estimated cost of the job. readOnly: true optimizerWeightDecay: type: number format: float description: Weight decay (L2 regularization) for optimizer. purpose: $ref: "#/components/schemas/gatewayPurpose" description: Scheduling purpose for this job. encryptionState: $ref: "#/components/schemas/gatewayEncryptionState" description: CMEK encryption state (authoritative, stamped at creation). readOnly: true required: - dataset gatewaySyncRlorTrainerJobCheckpointResponse: type: object properties: checkpoint: $ref: "#/components/schemas/gatewayRlorTrainerJobCheckpoint" description: >- The checkpoint after the sync attempt, including the availability classes known to the control plane. A successful regional-to-durable sync normally returns both SINGLE_REGION and MANAGED_DURABLE availability; a successful durable-to-regional sync normally returns MANAGED_DURABLE plus SINGLE_REGION for the requested region. gatewayTEFTDetails: type: object gatewayTimeSeries: type: object properties: labels: type: object additionalProperties: type: string title: >- Labels for this time series (e.g. {"deployment": "deploy1", "instance": "pod1"}) values: type: array items: type: object $ref: "#/components/schemas/gatewayTimeSeriesPoint" title: Array of [timestamp, value] pairs for this series gatewayTimeSeriesPoint: type: object properties: timestamp: type: string format: int64 title: Unix timestamp in seconds value: type: string title: The value at this timestamp gatewayTrail: type: object properties: name: type: string title: "Resource name: accounts/{AccountId}/trails/{TrailId}" readOnly: true displayName: type: string title: Human-readable display name for the trail createTime: type: string format: date-time title: Creation timestamp readOnly: true updateTime: type: string format: date-time title: Last update timestamp readOnly: true description: type: string title: Optional description of what this trail is tracking createdBy: type: string title: The email address of the user who created this trail readOnly: true langfuseConfig: type: string description: >- Langfuse configuration for this trail, inherited from the account at creation time. Contains API credentials for a Project within the customer's Langfuse Organization. Serialized JSON format: {"public_key":"pk-lf-xxx","secret_key":"sk-lf-xxx","host":"https://langfuse-prod.fireworks.ai"} If empty, the default Langfuse Organization/Project will be used. readOnly: true defaultModel: type: string description: >- Default model for requests through this trail. Format: "{provider}/{model_id}" (e.g., "anthropic/claude-3-5-sonnet-20240620"). Can be overridden per request. providerKey: type: string description: >- Provider API key for this trail. When creating a trail: provide the raw API key (e.g., "sk-ant-api03-xxxx...") After creation: this field contains a secret reference (e.g., "accounts/{account_id}/secrets/trail-xxx-provider-key") The LiteLLM gateway retrieves the actual key from Secret Manager using this reference. Can be overridden by specifying api_key in the request body. gatewayTrailMetrics: type: object properties: totalTraceCount: type: string format: int64 description: Total number of traces collected for this trail. readOnly: true firstTraceTime: type: string format: date-time description: Timestamp of the first trace in this trail. readOnly: true lastTraceTime: type: string format: date-time description: Timestamp of the most recent trace in this trail. readOnly: true promptTokenCount: type: string format: int64 description: Number of prompt tokens across all traces in this trail. readOnly: true completionTokenCount: type: string format: int64 description: Number of completion tokens across all traces in this trail. readOnly: true totalTokenCount: type: string format: int64 description: Number of total tokens across all traces in this trail. readOnly: true description: Metrics for a trail, fetched from the tracing service. gatewayTrainerShardingScheme: type: object properties: tensorParallelism: type: integer format: int32 description: Tensor-parallel degree. 0 means unspecified (server defaults to 1). pipelineParallelism: type: integer format: int32 description: >- Pipeline-parallel degree. 0 means unspecified (server defaults to 1). contextParallelism: type: integer format: int32 description: Context-parallel degree. 0 means unspecified (server defaults to 1). expertParallelism: type: integer format: int32 description: Expert-parallel degree. 0 means unspecified (server defaults to 1). sequenceParallelism: type: boolean description: Whether sequence parallelism should be enabled. description: Structured parallelism/sharding profile used by trainer launches. gatewayTrainingCurve: type: object properties: metricKey: type: string description: Metric key (e.g. "reward", "loss"). Natural key per session. steps: type: array items: type: string format: int64 description: Step values; aligned by index with `values`. values: type: array items: type: number format: double description: Scalar observations; aligned by index with `steps`. description: |- TrainingCurve is a contiguous time series of (step, value) pairs aligned by index, for a single metric on a TrainingSession. Item type in the AIP-132 list response for ListTrainingSessionTrainingCurves. gatewayTrainingRun: type: object properties: name: type: string readOnly: true createTime: type: string format: date-time readOnly: true updateTime: type: string format: date-time readOnly: true createdBy: type: string readOnly: true baseModel: type: string description: >- Base model resource name the run fine-tunes (accounts/{a}/models/{m}). Set once at create_model and immutable thereafter. loraRank: type: integer format: int32 description: LoRA rank for this run. Set once at create_model. status: $ref: "#/components/schemas/gatewayStatus" readOnly: true sourceTrainingSessionId: type: string description: >- The TrainingSession this run was first created under (its source session). A run is created under a session but persists across sessions — it can later be loaded under a different session — so this records the originating session for lineage. Set once at create and immutable. state: $ref: "#/components/schemas/gatewayTrainingRunState" readOnly: true description: >- TrainingRun is the durable, account-owned model lineage created by the Tinker SDK's create_lora_training_client. A TrainingSession (the ServiceClient connection) may host many runs, and a single run may span sessions across resume. gatewayTrainingRunState: type: string enum: - STATE_UNSPECIFIED - PENDING_MATERIALIZATION - READY - FAILED - EXPIRED default: STATE_UNSPECIFIED description: >- Materialization state for the trainer-side model backing this run. UNSPECIFIED is reserved for legacy rows created before the explicit state machine existed; reconcilers must not sweep it. - EXPIRED: Clean lifecycle expiry (owning session went idle / was torn down), not a trainer failure. Distinct from FAILED so callers can tell the two apart. gatewayTrainingSession: type: object properties: name: type: string readOnly: true displayName: type: string description: >- Human-readable display name of the training session. e.g. "Training session" Must be fewer than 64 characters long. createTime: type: string format: date-time readOnly: true updateTime: type: string format: date-time readOnly: true state: $ref: "#/components/schemas/gatewayTrainingSessionState" readOnly: true status: $ref: "#/components/schemas/gatewayStatus" readOnly: true createdBy: type: string readOnly: true referenceState: $ref: "#/components/schemas/TrainingSessionReferenceState" readOnly: true baseModel: type: string description: >- Base model resource name (accounts/{a}/models/{m}). Supplied at create_model (bind), not at create_session (reserve) — so OPTIONAL at the wire level and set-once / IMMUTABLE thereafter. description: >- TrainingSession represents a single training context on a shared trainer. It starts with a base model reference and can have LoRA adapters loaded or promoted via LoadTrainingSessionState. Each session gets its own inference route in the API gateway. gatewayTrainingSessionCheckpoint: type: object properties: name: type: string readOnly: true checkpointType: $ref: "#/components/schemas/gatewayCheckpointType" description: |- The kind of artifact — TRAINING_LORA for resumable training state, INFERENCE_LORA for an adapter that can be promoted to a Model. readOnly: true checkpointName: type: string description: The name given to this checkpoint when it was saved (e.g. "step-5"). readOnly: true createTime: type: string format: date-time readOnly: true promotable: type: boolean description: >- True when this checkpoint can be promoted to an inference Model (inference adapters). False for resumable training-state checkpoints. readOnly: true updateTime: type: string format: date-time description: >- Mirrors create_time. Checkpoints are immutable once saved, so there is no separate update event; present for resource-shape uniformity. readOnly: true availability: type: array items: type: object $ref: "#/components/schemas/gatewayCheckpointAvailability" description: |- Where the checkpoint's bytes are available — at most one entry per (availability_class, region). MANAGED_DURABLE for the durable copy; SINGLE_REGION for a region-local copy. readOnly: true description: >- A checkpoint saved during a TrainingSession. Returned by ListTrainingSessionCheckpoints and, when promotable, can be turned into a customer-owned Model via PromoteTrainingSessionCheckpoint. gatewayTrainingSessionChunkOutput: type: object properties: name: type: string readOnly: true step: type: integer format: int32 endOfEpoch: type: boolean outputDatasetName: type: string checkpointPath: type: string description: >- Server-signed URL (24h TTL) suitable for direct browser download. Empty string when signing is unavailable for the underlying object. epochStreamlogPath: type: string description: Server-signed URL (24h TTL). Empty when signing is unavailable. distribution: $ref: "#/components/schemas/gatewayDistributionSnapshot" epochId: type: string createTime: type: string format: date-time readOnly: true updateTime: type: string format: date-time readOnly: true gatewayTrainingSessionState: type: string enum: - TRAINING_SESSION_STATE_UNSPECIFIED - READY - FAILED - DELETING - EXPIRED default: TRAINING_SESSION_STATE_UNSPECIFIED description: |- AIP-216 lifecycle. DELETING is observable when DeleteTrainingSession releases the shared-trainer binding before the row is removed; Create is synchronous today and lands in READY directly, so no CREATING value is declared until an async Create path actually emits it (minimal-surface rule). gatewayTrainingShape: type: object properties: name: type: string title: >- The resource name of the training shape. e.g. accounts/my-account/trainingShapes/my-training-shape readOnly: true displayName: type: string description: >- Human-readable display name of the training shape. e.g. "Llama3 70B H200 BF16" Must be fewer than 64 characters long. description: type: string description: >- The description of the training shape. Must be fewer than 1000 characters long. createTime: type: string format: date-time description: The creation time of the training shape. readOnly: true updateTime: type: string format: date-time description: The update time for the training shape. readOnly: true baseModel: type: string title: The base model name. e.g. accounts/fireworks/models/llama-3-70b deploymentShapeVersion: type: string title: >- The validated deployment shape version used for numerics verification. e.g. accounts/fireworks/deploymentShapes/rft-qwen3-4b/versions/abcd1234 trainerImageTag: type: string description: >- The validated trainer runtime image tag used for numerics verification. trainerMode: $ref: "#/components/schemas/TrainingShapeTrainerMode" description: Trainer mode used for the validated launch profile. nodeCount: type: integer format: int32 description: Node count validated for the launch profile. trainerShardingScheme: $ref: "#/components/schemas/gatewayTrainerShardingScheme" description: >- Structured sharding/parallelism profile validated for the trainer launch. modelType: type: string description: The model type of the base model (e.g. "llama", "qwen"). readOnly: true parameterCount: type: string format: int64 description: >- The parameter count of the base model, used for billion-parameter bucket matching. readOnly: true acceleratorType: $ref: "#/components/schemas/gatewayAcceleratorType" description: The type of accelerator to use. acceleratorCount: type: integer format: int32 description: Total number of accelerators used by the job. baseModelWeightPrecision: $ref: "#/components/schemas/gatewayWeightPrecision" description: >- Default precision for base weights during training (prefer BF16; QLoRA/QAT omitted). maxSupportedContextLength: type: integer format: int32 description: Capacity limits validated for this shape. minTotalAcceleratorCount: type: integer format: int32 description: >- Minimum total accelerator count (world size, across all data-parallel replicas) this shape has been validated at. Written from shape-validation CI evidence; do not hand-edit downward. title: >- A training shape standardizes hardware/precision/parallelism defaults for training jobs of a given base model family and parameter bucket. required: - baseModel - trainerImageTag gatewayTrainingShapeVersion: type: object properties: name: type: string title: >- The resource name, e.g. accounts/.../trainingShapes/.../versions/{version_id} readOnly: true createTime: type: string format: date-time title: Creation time (used for ordering desc by default) readOnly: true snapshot: $ref: "#/components/schemas/gatewayTrainingShape" description: Full snapshot of the TrainingShape at this version. readOnly: true validated: type: boolean description: |- Whether this version has been validated through capacity tests. Only superusers can set this flag. public: type: boolean description: If true, this version will be publicly readable. latestValidated: type: boolean description: >- If true, this version is the latest validated version (at most one per shape). readOnly: true updateTime: type: string format: date-time description: Last update time of mutable fields such as validated/public. readOnly: true title: >- A training shape version is a specific immutable snapshot of a training shape. Versions are immutable, created on updates, and deleted when the parent training shape is deleted. gatewayTransformed: type: object properties: sourceDatasetId: type: string filter: type: string originalFormat: $ref: "#/components/schemas/DatasetFormat" required: - sourceDatasetId gatewayUploadDatasetResponse: type: object properties: filenameToSignedUrls: type: object additionalProperties: type: string title: Signed URLs for uploading dataset files gatewayUsageBucket: type: object properties: startTime: type: string format: date-time description: Start of the bucket. endTime: type: string format: date-time description: End of the bucket. lineItems: type: array items: type: object $ref: "#/components/schemas/gatewayLineItem" description: >- Line items for a single time bucket. The bucket size matches the request's granularity (one day for DAILY, the only granularity supported today; coarser buckets such as monthly may be added later). gatewayUser: type: object properties: name: type: string title: >- The resource name of the user. e.g. accounts/my-account/users/my-user readOnly: true displayName: type: string description: |- Human-readable display name of the user. e.g. "Alice" Must be fewer than 64 characters long. serviceAccount: type: boolean title: Whether this user is a service account (can only be set by admins) createTime: type: string format: date-time description: The creation time of the user. readOnly: true role: type: string description: >- The user's role: admin, user, contributor, inference-user, or custom. When set to "custom", the user's permissions are governed by permission_preset. email: type: string description: The user's email address. state: $ref: "#/components/schemas/gatewayUserState" description: The state of the user. readOnly: true status: $ref: "#/components/schemas/gatewayStatus" description: Contains information about the user status. readOnly: true updateTime: type: string format: date-time description: The update time for the user. readOnly: true permissionPreset: type: string description: >- The permission preset for this user. Only valid when role is "custom". required: - role gatewayUserState: type: string enum: - STATE_UNSPECIFIED - CREATING - READY - UPDATING - DELETING default: STATE_UNSPECIFIED gatewayUserUploaded: type: object gatewayUserUsageLimitIncreaseRequest: type: object properties: name: type: string title: >- Resource name. Format: accounts/{account_id}/usageLimitIncreaseRequests/{request_id} readOnly: true userId: type: string description: >- The user the request is for. Self-serve only: a member may request only for themselves, so this is also the requester. requestedLimit: $ref: "#/components/schemas/typeMoney" description: Requested per-user cap. Must exceed the current effective cap. requestState: $ref: "#/components/schemas/UserUsageLimitIncreaseRequestRequestState" description: Current state; server-owned (set on create and admin decision). readOnly: true createTime: type: string format: date-time description: When the request was filed. readOnly: true previousLimit: $ref: "#/components/schemas/typeMoney" description: The user's effective cap when the request was filed. readOnly: true decisionTime: type: string format: date-time description: When the request was approved or rejected. Unset while PENDING. readOnly: true decidedBy: type: string description: >- The admin who approved or rejected it (user id, else email). Unset while PENDING. readOnly: true description: >- UserUsageLimitIncreaseRequest is a member's request to raise their per-user serverless cap; an admin approves (applies the override) or rejects it. Full history is kept: one row per request. required: - userId - requestedLimit gatewayUserUsageLimits: type: object properties: name: type: string title: |- Resource name of the user usage-limits singleton. Format: accounts/{account_id}/users/{user_id}/usageLimits readOnly: true limitOverride: $ref: "#/components/schemas/typeMoney" description: The user's override cap. Unset means inherit the account default. effectiveLimit: $ref: "#/components/schemas/typeMoney" description: >- The user's resolved effective cap: their override, else the account default. Unset means the user has no cap (unlimited). Computed at read time; not stored. readOnly: true used: $ref: "#/components/schemas/typeMoney" description: Current-period usage. Computed at read time; not stored. readOnly: true enforcementMode: $ref: "#/components/schemas/gatewayAccountUsageLimitsEnforcementMode" description: >- Enforcement mode from the governing account policy. Read from the account usage-limits at read time; not stored on the user row. readOnly: true exceededUntil: type: string format: date-time description: >- When set, the user is over their limit and blocked until this time (the end of the current consumption period). Unset when the user is not over limit. Computed at read time; not stored. readOnly: true updateTime: type: string format: date-time description: >- The time this user's override was last set or changed. Unset when the user has no override (they inherit the account default); it tracks the override row, not account-default changes. readOnly: true description: >- UserUsageLimits is a single user's usage-limits singleton: their optional override cap, their resolved effective cap, their current-period usage, and whether (and until when) they are over their limit. There is exactly one per user. The override cap is stored; the effective cap, usage, enforcement mode, and exceeded_until are computed at read time. gatewayValidateAssertionsResponse: type: object properties: status: type: string metricToErrors: type: object additionalProperties: $ref: >- #/components/schemas/ValidateAssertionsResponseValidateAssertionError title: >- Response for validating assertions. The metric_to_errors contains all metrics from request, and the error_messages contains all errors for the metric code. (could be empty which means no errors) gatewayValidateModelConfigRequest: type: object properties: configJson: type: string description: The config JSON of the model. tokenizerConfigJson: type: string description: The tokenizer config JSON of the model. required: - configJson gatewayValidateModelUploadResponse: type: object properties: warnings: type: array items: type: string title: Warnings generated during validation (e.g., unknown config fields) gatewayWandbConfig: type: object properties: enabled: type: boolean description: Whether to enable wandb logging. apiKey: type: string description: The API key for the wandb service. project: type: string description: The project name for the wandb service. entity: type: string description: The entity name for the wandb service. runId: type: string description: The run ID for the wandb service. url: type: string description: The URL for the wandb service. readOnly: true description: >- WandbConfig is the configuration for the Weights & Biases (wandb) logging which will be used by a training job. gatewayWeightPrecision: type: string enum: - WEIGHT_PRECISION_UNSPECIFIED - BFLOAT16 - INT8 - NF4 - FP8 - FP4_FP8 default: WEIGHT_PRECISION_UNSPECIFIED description: |- The weight precision for model training/inference. - BFLOAT16: no quantization applied - INT8: enable 8-bit quantization with LLM.int8() - NF4: enable 4-bit quantization with LLM.nf4() - FP8: base model quantization in FP8 - FP4_FP8: base model linear module quantization in FP4, mixed with experts and some special keys/layers in FP8. gatewayWeightedRandomStrategy: type: object description: Use replica count as weight. googlelongrunningOperation: type: object properties: name: type: string description: >- The server-assigned name, which is only unique within the same service that originally returns it. If you use the default HTTP mapping, the `name` should be a resource name ending with `operations/{unique_id}`. metadata: $ref: "#/components/schemas/protobufAny" description: >- Service-specific metadata associated with the operation. It typically contains progress information and common metadata such as create time. Some services might not provide such metadata. Any method that returns a long-running operation should document the metadata type, if any. done: type: boolean description: >- If the value is `false`, it means the operation is still in progress. If `true`, the operation is completed, and either `error` or `response` is available. error: $ref: "#/components/schemas/googlerpcStatus" description: >- The error result of the operation in case of failure or cancellation. response: $ref: "#/components/schemas/protobufAny" description: >- The normal, successful response of the operation. If the original method returns no data on success, such as `Delete`, the response is `google.protobuf.Empty`. If the original method is standard `Get`/`Create`/`Update`, the response should be the resource. For other methods, the response should have the type `XxxResponse`, where `Xxx` is the original method name. For example, if the original method name is `TakeSnapshot()`, the inferred response type is `TakeSnapshotResponse`. description: >- This resource represents a long-running operation that is the result of a network API call. googlerpcStatus: type: object properties: code: type: integer format: int32 description: |- The status code, which should be an enum value of [google.rpc.Code][google.rpc.Code]. message: type: string description: >- A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the [google.rpc.Status.details][google.rpc.Status.details] field, or localized by the client. details: type: array items: type: object $ref: "#/components/schemas/protobufAny" description: >- A list of messages that carry the error details. There is a common set of message types for APIs to use. description: >- The `Status` type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by [gRPC](https://github.com/grpc). Each `Status` message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the [API Design Guide](https://cloud.google.com/apis/design/errors). protobufAny: type: object properties: "@type": type: string description: >- A URL/resource name that uniquely identifies the type of the serialized protocol buffer message. This string must contain at least one "/" character. The last segment of the URL's path must represent the fully qualified name of the type (as in `path/google.protobuf.Duration`). The name should be in a canonical form (e.g., leading "." is not accepted). In practice, teams usually precompile into the binary all types that they expect it to use in the context of Any. However, for URLs which use the scheme `http`, `https`, or no scheme, one can optionally set up a type server that maps type URLs to message definitions as follows: * If no scheme is provided, `https` is assumed. * An HTTP GET on the URL must yield a [google.protobuf.Type][] value in binary format, or produce an error. * Applications are allowed to cache lookup results based on the URL, or have them precompiled into a binary to avoid any lookup. Therefore, binary compatibility needs to be preserved on changes to types. (Use versioned type names to manage breaking changes.) Note: this functionality is not currently available in the official protobuf release, and it is not used for type URLs beginning with type.googleapis.com. Schemes other than `http`, `https` (or the empty scheme) might be used with implementation specific semantics. additionalProperties: {} description: >- `Any` contains an arbitrary serialized protocol buffer message along with a URL that describes the type of the serialized message. Protobuf library provides support to pack/unpack Any values in the form of utility functions or additional generated methods of the Any type. Example 1: Pack and unpack a message in C++. Foo foo = ...; Any any; any.PackFrom(foo); ... if (any.UnpackTo(&foo)) { ... } Example 2: Pack and unpack a message in Java. Foo foo = ...; Any any = Any.pack(foo); ... if (any.is(Foo.class)) { foo = any.unpack(Foo.class); } Example 3: Pack and unpack a message in Python. foo = Foo(...) any = Any() any.Pack(foo) ... if any.Is(Foo.DESCRIPTOR): any.Unpack(foo) ... Example 4: Pack and unpack a message in Go foo := &pb.Foo{...} any, err := anypb.New(foo) if err != nil { ... } ... foo := &pb.Foo{} if err := any.UnmarshalTo(foo); err != nil { ... } The pack methods provided by protobuf library will by default use 'type.googleapis.com/full.type.name' as the type URL and the unpack methods only use the fully qualified type name after the last '/' in the type URL, for example "foo.bar.com/x/y.z" will yield type name "y.z". JSON The JSON representation of an `Any` value uses the regular representation of the deserialized, embedded message, with an additional field `@type` which contains the type URL. Example: package google.profile; message Person { string first_name = 1; string last_name = 2; } { "@type": "type.googleapis.com/google.profile.Person", "firstName": , "lastName": } If the embedded message type is well-known and has a custom JSON representation, that representation will be embedded adding a field `value` which holds the custom JSON in addition to the `@type` field. Example (for message [google.protobuf.Duration][]): { "@type": "type.googleapis.com/google.protobuf.Duration", "value": "1.212s" } protobufNullValue: type: string enum: - NULL_VALUE default: NULL_VALUE description: >- `NullValue` is a singleton enumeration to represent the null value for the `Value` type union. The JSON representation for `NullValue` is JSON `null`. - NULL_VALUE: Null value. typeDate: type: object properties: year: type: integer format: int32 description: >- Year of the date. Must be from 1 to 9999, or 0 to specify a date without a year. month: type: integer format: int32 description: >- Month of a year. Must be from 1 to 12, or 0 to specify a year without a month and day. day: type: integer format: int32 description: >- Day of a month. Must be from 1 to 31 and valid for the year and month, or 0 to specify a year by itself or a year and month where the day isn't significant. description: >- * A full date, with non-zero year, month, and day values * A month and day value, with a zero year, such as an anniversary * A year on its own, with zero month and day values * A year and month value, with a zero day, such as a credit card expiration date Related types are [google.type.TimeOfDay][google.type.TimeOfDay] and `google.protobuf.Timestamp`. title: >- Represents a whole or partial calendar date, such as a birthday. The time of day and time zone are either specified elsewhere or are insignificant. The date is relative to the Gregorian Calendar. This can represent one of the following: typeMoney: type: object properties: currencyCode: type: string description: The three-letter currency code defined in ISO 4217. units: type: string format: int64 description: >- The whole units of the amount. For example if `currencyCode` is `"USD"`, then 1 unit is one US dollar. nanos: type: integer format: int32 description: >- Number of nano (10^-9) units of the amount. The value must be between -999,999,999 and +999,999,999 inclusive. If `units` is positive, `nanos` must be positive or zero. If `units` is zero, `nanos` can be positive, zero, or negative. If `units` is negative, `nanos` must be negative or zero. For example $-1.75 is represented as `units`=-1 and `nanos`=-750,000,000. description: Represents an amount of money with its currency type. FileUploadResponse: type: object properties: id: type: string description: The dataset id. object: type: string description: The object type, which is always file. bytes: type: integer format: int64 description: The size of the file, in bytes. created_at: type: integer format: int64 description: The Unix timestamp (in seconds) for when the file was created. filename: type: string description: The name of the file. purpose: type: string description: The intended purpose of the file. CreateResponse: properties: model: type: string title: Model description: >- The model to use for generating the response. Example: `accounts//models/`. input: anyOf: - type: string - items: additionalProperties: true type: object type: array title: Input description: >- The input to the model. Can be a simple text string or a list of message objects for complex inputs with multiple content types. previous_response_id: anyOf: - type: string - type: "null" title: Previous Response Id description: >- The ID of a previous response to continue the conversation from. When provided, the conversation history from that response will be automatically loaded. instructions: anyOf: - type: string - type: "null" title: Instructions description: >- System instructions that guide the model's behavior throughout the conversation. Similar to a system message. max_output_tokens: anyOf: - type: integer - type: "null" title: Max Output Tokens description: >- The maximum number of tokens that can be generated in the response. Must be at least 1. If not specified, the model will generate up to its maximum context length. max_tool_calls: anyOf: - type: integer minimum: 1 - type: "null" title: Max Tool Calls description: >- The maximum number of tool calls allowed in a single response. Useful for controlling costs and limiting tool execution. Must be at least 1. metadata: anyOf: - additionalProperties: true type: object - type: "null" title: Metadata description: >- Set of up to 16 key-value pairs that can be attached to the response. Useful for storing additional information in a structured format. parallel_tool_calls: anyOf: - type: boolean - type: "null" title: Parallel Tool Calls description: >- Whether to enable parallel function calling during tool use. When true, the model can call multiple tools simultaneously. Default is True. default: true reasoning: anyOf: - additionalProperties: true type: object - type: "null" title: Reasoning description: >- Configuration for reasoning output. When enabled, the model will return its reasoning process along with the response. store: anyOf: - type: boolean - type: "null" title: Store description: >- Whether to store the response. When set to false, the response will not be stored and will not be retrievable via the API. This is useful for ephemeral or sensitive data. See an example in our [Controlling Response Storage cookbook](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/mcp_server_with_store_false_argument.ipynb). Default is True. default: true stream: anyOf: - type: boolean - type: "null" title: Stream description: >- Whether to stream the response back as Server-Sent Events (SSE). When true, tokens are sent incrementally as they are generated. Default is False. default: false temperature: anyOf: - type: number maximum: 2 minimum: 0 - type: "null" title: Temperature description: >- The sampling temperature to use, between 0 and 2. Higher values like 0.8 make output more random, while lower values like 0.2 make it more focused and deterministic. Default is 1.0. default: 1 text: anyOf: - additionalProperties: true type: object - type: "null" title: Text description: >- Text generation configuration parameters. Used for advanced text generation settings. tool_choice: anyOf: - type: string - additionalProperties: true type: object - type: "null" title: Tool Choice description: >- Controls which (if any) tool the model should use. Can be 'none' (never call tools), 'auto' (model decides), 'required' (must call at least one tool), or an object specifying a particular tool to call. Default is 'auto'. default: auto tools: anyOf: - items: additionalProperties: true type: object type: array - type: "null" title: Tools description: >- A list of MCP tools the model may call. See our cookbooks for examples on [basic MCP usage](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/fireworks_mcp_examples.ipynb) and [streaming with MCP](https://github.com/fw-ai/cookbook/blob/main/archived/learn/response-api/fireworks_mcp_with_streaming.ipynb). top_p: anyOf: - type: number maximum: 1 minimum: 0 - type: "null" title: Top P description: >- An alternative to temperature sampling, called nucleus sampling, where the model considers the results of tokens with top_p probability mass. So 0.1 means only tokens comprising the top 10% probability mass are considered. Default is 1.0. We generally recommend altering this or temperature but not both. default: 1 truncation: anyOf: - type: string - type: "null" title: Truncation description: >- The truncation strategy to use for the context when it exceeds the model's maximum length. Can be 'auto' (automatically truncate) or 'disabled' (return error if context too long). Default is 'disabled'. default: disabled user: anyOf: - type: string - type: "null" title: User description: >- A unique identifier representing your end-user, which can help Fireworks to monitor and detect abuse. This can be a username, email, or any other unique identifier. type: object required: - model - input title: CreateResponse description: >- Request model for creating a new response. This model defines all the parameters needed to create a new model response, including model configuration, input data, tool definitions, and conversation continuation. DeleteResponse: properties: message: type: string title: Message description: Confirmation message example: Response deleted successfully type: object required: - message title: DeleteResponse description: Response model for deleting a response. HTTPValidationError: properties: detail: items: $ref: "#/components/schemas/ValidationError" type: array title: Detail type: object title: HTTPValidationError Message: properties: id: type: string title: Id description: The unique identifier of the message. type: type: string title: Type description: The object type, always 'message'. default: message role: type: string title: Role description: >- The role of the message sender. Can be 'user', 'assistant', or 'system'. content: items: $ref: "#/components/schemas/MessageContent" type: array title: Content description: >- An array of content parts that make up the message. Each part has a type and associated data. status: type: string title: Status description: The status of the message. Can be 'in_progress' or 'completed'. type: object required: - id - role - content - status title: Message description: Represents a message in a conversation. MessageContent: properties: type: type: string title: Type description: >- The type of the content part. Can be 'input_text', 'output_text', 'image', etc. text: anyOf: - type: string - type: "null" title: Text description: The text content, if applicable. type: object required: - type title: MessageContent description: Represents a piece of content within a message. Response: properties: id: anyOf: - type: string - type: "null" title: Id description: The unique identifier of the response. Will be None if store=False. object: type: string title: Object description: The object type, which is always 'response'. default: response created_at: type: integer title: Created At description: The Unix timestamp (in seconds) when the response was created. status: type: string title: Status description: >- The status of the response. Can be 'completed', 'in_progress', 'incomplete', 'failed', or 'cancelled'. model: type: string title: Model description: >- The model used to generate the response (e.g., `accounts//models/`). output: items: anyOf: - $ref: "#/components/schemas/Message" - $ref: "#/components/schemas/ToolCall" - $ref: "#/components/schemas/ToolOutput" type: array title: Output description: >- An array of output items produced by the model. Can contain messages, tool calls, and tool outputs. previous_response_id: anyOf: - type: string - type: "null" title: Previous Response Id description: >- The ID of the previous response in the conversation, if this response continues a conversation. usage: anyOf: - additionalProperties: true type: object - type: "null" title: Usage description: >- Token usage information for the request. Contains 'prompt_tokens', 'completion_tokens', and 'total_tokens'. error: anyOf: - additionalProperties: true type: object - type: "null" title: Error description: >- Error information if the response failed. Contains 'type', 'code', and 'message' fields. incomplete_details: anyOf: - additionalProperties: true type: object - type: "null" title: Incomplete Details description: >- Details about why the response is incomplete, if status is 'incomplete'. Contains 'reason' field which can be 'max_output_tokens', 'max_tool_calls', or 'content_filter'. instructions: anyOf: - type: string - type: "null" title: Instructions description: >- System instructions that guide the model's behavior. Similar to a system message. max_output_tokens: anyOf: - type: integer - type: "null" title: Max Output Tokens description: >- The maximum number of tokens that can be generated in the response. Must be at least 1. max_tool_calls: anyOf: - type: integer minimum: 1 - type: "null" title: Max Tool Calls description: >- The maximum number of tool calls allowed in a single response. Must be at least 1. parallel_tool_calls: type: boolean title: Parallel Tool Calls description: >- Whether to enable parallel function calling during tool use. Default is True. default: true reasoning: anyOf: - additionalProperties: true type: object - type: "null" title: Reasoning description: >- Reasoning output from the model, if reasoning is enabled. Contains 'content' and 'type' fields. store: anyOf: - type: boolean - type: "null" title: Store description: >- Whether to store this response for future retrieval. If False, the response will not be persisted and previous_response_id cannot reference it. Default is True. default: true temperature: type: number maximum: 2 minimum: 0 title: Temperature description: >- The sampling temperature to use, between 0 and 2. Higher values like 0.8 make output more random, while lower values like 0.2 make it more focused and deterministic. Default is 1.0. default: 1 text: anyOf: - additionalProperties: true type: object - type: "null" title: Text description: Text generation configuration parameters, if applicable. tool_choice: anyOf: - type: string - additionalProperties: true type: object title: Tool Choice description: >- Controls which (if any) tool the model should use. Can be 'none', 'auto', 'required', or an object specifying a particular tool. Default is 'auto'. default: auto tools: items: additionalProperties: true type: object type: array title: Tools description: >- A list of tools the model may call. Each tool is defined with a type and function specification following the OpenAI tool format. Supports 'function', 'mcp', 'sse', and 'python' tool types. top_p: type: number maximum: 1 minimum: 0 title: Top P description: >- An alternative to temperature sampling, called nucleus sampling, where the model considers the results of tokens with top_p probability mass. So 0.1 means only tokens comprising the top 10% probability mass are considered. Default is 1.0. default: 1 truncation: type: string title: Truncation description: >- The truncation strategy to use for the context. Can be 'auto' or 'disabled'. Default is 'disabled'. default: disabled user: anyOf: - type: string - type: "null" title: User description: >- A unique identifier representing your end-user, which can help Fireworks to monitor and detect abuse. metadata: anyOf: - additionalProperties: true type: object - type: "null" title: Metadata description: >- Set of up to 16 key-value pairs that can be attached to the response. Useful for storing additional information about the response in a structured format. type: object required: - created_at - status - model - output title: Response description: >- Represents a response object returned from the API. A response includes the model output, token usage, configuration parameters, and metadata about the conversation state. ResponseList: properties: object: type: string title: Object description: The object type, which is always 'list'. default: list data: items: $ref: "#/components/schemas/Response" type: array title: Data description: >- An array of response objects, sorted by creation time in descending order (most recent first). has_more: type: boolean title: Has More description: >- Indicates whether there are more responses available beyond this page. If true, use the 'last_id' value as the 'after' cursor to fetch the next page. first_id: anyOf: - type: string - type: "null" title: First Id description: >- The ID of the first response in the current page. Used for pagination. last_id: anyOf: - type: string - type: "null" title: Last Id description: >- The ID of the last response in the current page. Use this as the 'after' cursor to fetch the next page if has_more is true. type: object required: - data - has_more title: ResponseList description: |- Response model for listing responses. Returned from the GET /v1/responses endpoint. Provides a paginated list of response objects with cursor-based pagination support. ToolCall: properties: id: type: string title: Id description: The unique identifier of the tool call. type: type: string title: Type description: The type of tool call. Can be 'function_call' or 'mcp_call'. call_id: anyOf: - type: string - type: "null" title: Call Id description: >- The call ID for function calls, used to match with function_call_output. name: anyOf: - type: string - type: "null" title: Name description: The name of the function to call (for function_call type). arguments: anyOf: - type: string - type: "null" title: Arguments description: >- The arguments for the function call as a JSON string (for function_call type). status: anyOf: - type: string - type: "null" title: Status description: >- The status of the tool call. Can be 'in_progress', 'completed', or 'incomplete'. function: anyOf: - additionalProperties: true type: object - type: "null" title: Function description: >- The function definition for function tool calls. Contains 'name' and 'arguments' keys. Deprecated for function_call type. mcp: anyOf: - additionalProperties: true type: object - type: "null" title: Mcp description: >- The MCP (Model Context Protocol) tool call definition for MCP tool calls. type: object required: - id - type title: ToolCall description: Represents a tool call made by the model. ToolOutput: properties: type: type: string title: Type description: The object type, always 'tool_output'. default: tool_output tool_call_id: type: string title: Tool Call Id description: The ID of the tool call that this output corresponds to. output: type: string title: Output description: The output content from the tool execution. type: object required: - tool_call_id - output title: ToolOutput description: Represents the output/result of a tool call. ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type input: title: Input ctx: type: object title: Context type: object required: - loc - msg - type title: ValidationError AllowedToolsConfig: properties: mode: type: string enum: - auto - required title: Mode description: >- `auto` lets the model pick between generating a message and calling one of the allowed tools. `required` forces the model to call one of them. tools: items: additionalProperties: true type: object type: array title: Tools description: >- A list of tool references the model is allowed to invoke. Each entry must be of the form `{"type": "function", "function": {"name": "..."}}`. The referenced tools must also appear in the top-level `tools` list. additionalProperties: false type: object required: - mode - tools title: AllowedToolsConfig description: >- OpenAI-compatible allowed_tools sub-object. See https://platform.openai.com/docs/guides/function-calling#tool-choice The ``tools`` list mirrors the ``tools`` list passed at the top level: each entry must be ``{"type": "function", "function": {"name": ""}}``. Fireworks only supports function tools today, so other ``type`` values are rejected. AllowedToolsSelection: properties: type: type: string const: allowed_tools title: Type allowed_tools: $ref: "#/components/schemas/AllowedToolsConfig" additionalProperties: false type: object required: - type - allowed_tools title: AllowedToolsSelection description: >- OpenAI-compatible ``tool_choice.type = "allowed_tools"`` selection. Restricts the model's tool calls to a subset of the top-level ``tools`` list *without modifying the prompt*, preserving prompt cache hits. The full set of tool definitions is still rendered into the prompt for the model; only the grammar (and downstream parsing) is constrained to the allowed subset. ChatCompletionFunction: properties: name: type: string title: Name description: >- The name of the function to be called. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64. description: anyOf: - type: string - type: "null" title: Description description: >- A description of what the function does, used by the model to choose when and how to call the function. parameters: additionalProperties: true type: object title: Parameters description: >- The parameters the function accepts, described as a JSON Schema object. The JSON Schema object should have the following structure: ```json { "type": "object", "required": ["param1", "param2"], "properties": { "param1": { "type": "string", "description": "..." }, "param2": { "type": "number", "description": "..." } } } ``` - The `type` field must be `"object"`. - The `required` field is an array of strings indicating which parameters are required. - The `properties` field is a map of property names to their definitions, where each property is an object with `type` (string) and `description` (string) fields. To describe a function that accepts no parameters, provide the value: ```json {"type": "object", "properties": {}} ``` strict: anyOf: - type: boolean - type: "null" title: Strict additionalProperties: false type: object required: - name title: ChatCompletionFunction ChatCompletionMessageToolCall: properties: id: anyOf: - type: string - type: "null" title: Id description: The ID of the tool call. type: type: string title: Type description: The type of the tool. Currently, only `function` is supported. default: function function: anyOf: - $ref: "#/components/schemas/ChatCompletionMessageToolCallFunction" - type: string title: Function description: The function that the model called. additionalProperties: false type: object required: - function title: ChatCompletionMessageToolCall ChatCompletionMessageToolCallFunction: properties: name: anyOf: - type: string - type: "null" title: Name description: The name of the function to call. arguments: anyOf: - type: string - additionalProperties: true type: object - type: "null" title: Arguments description: >- The arguments to call the function with, as generated by the model in JSON format. Note that the model does not always generate valid JSON, and may hallucinate parameters not defined by your function schema. Validate the arguments in your code before calling your function. additionalProperties: false type: object title: ChatCompletionMessageToolCallFunction ChatCompletionRequest: properties: model: type: string title: Model description: |- The name of the model to use. Example: `"accounts/fireworks/models/kimi-k2-instruct-0905"` messages: items: $ref: "#/components/schemas/ChatMessage" type: array title: Messages description: >- A list of messages comprising the conversation so far. When ``prompt_token_ids`` is supplied this field is ignored (the caller is taking responsibility for chat-template rendering and tokenization upstream). tools: items: $ref: "#/components/schemas/ChatCompletionTool" type: array title: Tools description: >- A list of tools the model may call. Currently, only functions are supported as a tool. Use this to provide a list of functions the model may generate JSON inputs for. See the our [model library](https://app.fireworks.ai/models/?filter=LLM&functionCalling=true) for the list of supported models tool_choice: anyOf: - type: string enum: - auto - none - any - required - $ref: "#/components/schemas/FunctionSelection" - $ref: "#/components/schemas/AllowedToolsSelection" title: Tool Choice description: >- Controls which (if any) tool is called by the model. - `none`: the model will not call any tool and instead generates a message. - `auto`: the model can pick between generating a message or calling one or more tools. - `required` (alias: `any`): the model must call one or more tools. To force a specific function, pass an object of the form `{ "type": "function", "name": "my_function" }` or `{ "type": "function", "function": { "name": "my_function" } }` for OpenAI compatibility. - To restrict tool calls to a subset of the provided tools while keeping the full tool list in the prompt (preserving prompt cache hits), pass `{ "type": "allowed_tools", "allowed_tools": { "mode": "auto"|"required", "tools": [{ "type": "function", "function": { "name": "..." } }] } }`. Equivalent OpenAI Responses-style flat form is also accepted: `{ "type": "allowed_tools", "mode": "...", "tools": [...] }`. default: auto stream: anyOf: - type: boolean - type: "null" title: Stream description: >- Whether to stream back partial progress. If set, tokens will be sent as data-only [server-sent events](https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format) as they become available, with the stream terminated by a `data: [DONE]` message. default: false response_format: anyOf: - $ref: "#/components/schemas/ResponseFormat" - type: "null" description: >- Allows to force the model to produce specific output format. Setting to `{ "type": "json_object" }` enables JSON mode, which guarantees the message the model generates is valid JSON. If `"type"` is `"json_schema"`, a JSON schema must be provided. E.g., `response_format = {"type": "json_schema", "json_schema": }`. Important: when using JSON mode, it's crucial to also instruct the model to produce JSON via a system or user message. Without this, the model may generate an unending stream of whitespace until the generation reaches the token limit, resulting in a long-running and seemingly "stuck" request. Also note that the message content may be partially cut off if `finish_reason="length"`, which indicates the generation exceeded `max_tokens` or the conversation exceeded the max context length. In this case the return value might not be a valid JSON. temperature: anyOf: - type: number - type: "null" title: Temperature description: >- What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. We generally recommend altering this or top_p but not both. Required range: `0 <= x <= 2` Example: `1` top_k: anyOf: - type: integer - type: "null" title: Top K description: >- Top-k sampling is another sampling method where the k most probable next tokens are filtered and the probability mass is redistributed among only those k next tokens. The value of k controls the number of candidates for the next token at each step during text generation. Must be between 0 and 100. Required range: `0 <= x <= 100` Example: `50` user: anyOf: - type: string - type: "null" title: User description: >- A unique identifier representing your end-user, which can help monitor and detect abuse. prompt_cache_key: anyOf: - type: string - type: "null" title: Prompt Cache Key description: >- A key used for prompt caching session affinity. Requests with the same prompt_cache_key are routed to the same backend to maximize KV cache hit rates. This is the preferred field for session affinity (takes priority over the 'user' field). prompt_cache_isolation_key: anyOf: - type: string - type: "null" title: Prompt Cache Isolation Key description: Isolation key for prompt caching to separate cache entries. raw_output: anyOf: - type: boolean - type: "null" title: Raw Output description: Return raw output from the model. default: false perf_metrics_in_response: anyOf: - type: boolean - type: "null" title: Perf Metrics In Response description: >- Whether to include performance metrics in the response body. **Non-streaming requests:** Performance metrics are always included in response headers (e.g., `fireworks-prompt-tokens`, `fireworks-server-time-to-first-token`). Setting this to `true` additionally includes the same metrics in the response body under the `perf_metrics` field. **Streaming requests:** Performance metrics are only included in the response body under the `perf_metrics` field in the final chunk (when `finish_reason` is set). This is because headers may not be accessible during streaming. The response body `perf_metrics` field contains the following metrics: **Basic Metrics (all deployments):** - `prompt-tokens`: Number of tokens in the prompt - `cached-prompt-tokens`: Number of cached prompt tokens - `server-time-to-first-token`: Time from request start to first token (in seconds) - `server-processing-time`: Total processing time (in seconds, only for completed requests) **Predicted Outputs Metrics:** - `speculation-prompt-tokens`: Number of speculative prompt tokens - `speculation-prompt-matched-tokens`: Number of matched speculative prompt tokens (for completed requests) **Dedicated Deployment Only Metrics:** - `speculation-generated-tokens`: Number of speculative generated tokens (for completed requests) - `speculation-acceptance`: Speculation acceptance rates by position - `backend-host`: Hostname of the backend server - `num-concurrent-requests`: Number of concurrent requests - `deployment`: Deployment name - `tokenizer-queue-duration`: Time spent in tokenizer queue - `tokenizer-duration`: Time spent in tokenizer - `prefill-queue-duration`: Time spent in prefill queue - `prefill-duration`: Time spent in prefill - `generation-queue-duration`: Time spent in generation queue - `generation-duration`: Time spent in generation default: false stream_options: anyOf: - $ref: "#/components/schemas/StreamOptions" - type: "null" description: >- Options for streaming responses. Only valid with ``stream=true``. Fireworks includes a final SSE chunk carrying usage totals by default; set ``include_usage=false`` to opt out. The ``buffer_tokens`` / ``buffer_ms`` / ``buffer_mode`` fields coalesce SSE chunks by token count and/or time (vLLM backend; override the deployment default). "n": type: integer title: "N" description: >- How many completions to generate for each prompt. **Note:** Because this parameter generates many completions, it can quickly consume your token quota. Use carefully and ensure that you have reasonable settings for `max_tokens` and `stop`. Required range: `1 <= x <= 128` Example: `1` default: 1 service_tier: type: string enum: - auto - default - flex - priority title: Service Tier description: >- The service tier to use for the request. Specifies the processing type used for serving the request. Only "priority" is supported, while all other values will be treated as "default" tier. default: default stop: anyOf: - type: string - items: type: string type: array - type: "null" title: Stop description: >- Up to 4 sequences where the API will stop generating further tokens. The returned text will NOT contain the stop sequence. max_tokens: anyOf: - type: integer - type: "null" title: Max Tokens description: >- The maximum number of tokens to generate in the completion. If the token count of your prompt plus max_tokens exceeds the model's context length, the behavior depends on context_length_exceeded_behavior. By default, max_tokens will be lowered to fit in the context window instead of returning an error. max_completion_tokens: anyOf: - type: integer - type: "null" title: Max Completion Tokens description: Alias for max_tokens. Cannot be specified together with max_tokens. top_p: anyOf: - type: number - type: "null" title: Top P description: >- An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. We generally recommend altering this or temperature but not both. Required range: `0 <= x <= 1` Example: `1` min_p: anyOf: - type: number - type: "null" title: Min P description: >- Minimum probability threshold for token selection. Only tokens with probability >= min_p are considered for selection. This is an alternative to `top_p` and `top_k` sampling. Required range: `0 <= x <= 1` typical_p: anyOf: - type: number - type: "null" title: Typical P description: >- Typical-p sampling is an alternative to nucleus sampling. It considers the most typical tokens whose cumulative probability is at most typical_p. Required range: `0 <= x <= 1` frequency_penalty: anyOf: - type: number - type: "null" title: Frequency Penalty description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. Reasonable value is around 0.1 to 1 if the aim is to just reduce repetitive samples somewhat. If the aim is to strongly suppress repetition, then one can increase the coefficients up to 2, but this can noticeably degrade the quality of samples. Negative values can be used to increase the likelihood of repetition. See also `presence_penalty` for penalizing tokens that have at least one appearance at a fixed rate. OpenAI compatible (follows OpenAI's conventions for handling token frequency and repetition penalties). Required range: `-2 <= x <= 2` presence_penalty: anyOf: - type: number - type: "null" title: Presence Penalty description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Reasonable value is around 0.1 to 1 if the aim is to just reduce repetitive samples somewhat. If the aim is to strongly suppress repetition, then one can increase the coefficients up to 2, but this can noticeably degrade the quality of samples. Negative values can be used to increase the likelihood of repetition. See also `frequency_penalty` for penalizing tokens at an increasing rate depending on how often they appear. OpenAI compatible (follows OpenAI's conventions for handling token frequency and repetition penalties). Required range: `-2 <= x <= 2` repetition_penalty: anyOf: - type: number - type: "null" title: Repetition Penalty description: >- Applies a penalty to repeated tokens to discourage or encourage repetition. A value of `1.0` means no penalty, allowing free repetition. Values above `1.0` penalize repetition, reducing the likelihood of repeating tokens. Values between `0.0` and `1.0` reward repetition, increasing the chance of repeated tokens. For a good balance, a value of `1.2` is often recommended. Note that the penalty is applied to both the generated output and the prompt in decoder-only models. Required range: `0 <= x <= 2` mirostat_target: anyOf: - type: number - type: "null" title: Mirostat Target description: >- Defines the target perplexity for the Mirostat algorithm. Perplexity measures the unpredictability of the generated text, with higher values encouraging more diverse and creative outputs, while lower values prioritize predictability and coherence. The algorithm dynamically adjusts the token selection to maintain this target during text generation. If not specified, Mirostat sampling is disabled. mirostat_lr: anyOf: - type: number - type: "null" title: Mirostat Lr description: >- Specifies the learning rate for the Mirostat sampling algorithm, which controls how quickly the model adjusts its token distribution to maintain the target perplexity. A smaller value slows down the adjustments, leading to more stable but gradual shifts, while higher values speed up corrections at the cost of potential instability. seed: anyOf: - type: integer - type: "null" title: Seed description: Random seed for deterministic sampling. logprobs: anyOf: - type: integer - type: boolean - type: "null" title: Logprobs description: >- Include log probabilities in the response. This accepts either a boolean or an integer: If set to `true`, log probabilities are included and the number of alternatives can be controlled via `top_logprobs` (OpenAI-compatible behavior). If set to an integer N, include log probabilities for up to N most likely tokens per position in the legacy format. N must be between 0 and the deployment's `--max-logprobs` limit (5 by default). The API will always return the logprob of the sampled token, so there may be up to `logprobs+1` elements in the response when an integer is used. top_logprobs: anyOf: - type: integer - type: "null" title: Top Logprobs description: >- An integer specifying the number of most likely tokens to return at each token position, each with an associated log probability. Must be between 0 and the deployment's `--max-logprobs` limit (5 by default). When `logprobs` is set, `top_logprobs` can be used to modify how many top log probabilities are returned. If `top_logprobs` is not set, the API will return up to `logprobs` tokens per position. sampling_mask: anyOf: - type: string enum: - count - non_zero_list - non_zero_buffer - type: "null" title: Sampling Mask description: >- Opt-in sampling mask metadata for generated tokens. When set to `"count"`, each generated token in the new logprobs format includes the number of token logits still eligible for sampling after filters such as top_p and top_k are applied. `"non_zero_list"` additionally returns active token IDs in `sampling_mask`; `"non_zero_buffer"` additionally returns a base64-encoded little-endian uint32 buffer of active token IDs. Non-zero payloads are omitted for positions with more active tokens than 1000. echo: anyOf: - type: boolean - type: "null" title: Echo description: Echo back the prompt in addition to the completion. default: false echo_last: anyOf: - type: integer - type: "null" title: Echo Last description: >- Echo back the last N tokens of the prompt in addition to the completion. This is useful for obtaining logprobs of the prompt suffix but without transferring too much data. Passing `echo_last=len(prompt)` is the same as `echo=True` ignore_eos: type: boolean title: Ignore Eos description: >- This setting controls whether the model should ignore the End of Sequence (EOS) token. When set to `True`, the model will continue generating tokens even after the EOS token is produced. By default, it stops when the EOS token is reached. default: false context_length_exceeded_behavior: type: string enum: - error - truncate title: Context Length Exceeded Behavior description: >- What to do if the token count of prompt plus `max_tokens` exceeds the model's context window. Passing `truncate` limits the `max_tokens` to at most `context_window_length - prompt_length`. This is the default. Passing `error` would trigger a request error. The default of `'truncate'` is selected as it allows to ask for high `max_tokens` value while respecting the context window length without having to do client-side prompt tokenization. Note, that it differs from OpenAI's behavior that matches that of `error`. default: truncate logit_bias: anyOf: - additionalProperties: type: number type: object - type: "null" title: Logit Bias description: >- Modify the likelihood of specified tokens appearing in the completion. Accepts a json object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. speculation: anyOf: - type: string - items: type: integer type: array - type: "null" title: Speculation description: Speculative decoding prompt or token IDs to speed up generation. prediction: anyOf: - $ref: "#/components/schemas/PredictedOutput" - type: string - type: "null" title: Prediction description: >- OpenAI-compatible predicted output for speculative decoding. Can be a PredictedOutput object or a simple string. Automatically transformed to speculation. metadata: anyOf: - additionalProperties: type: string type: object - type: "null" title: Metadata description: >- Additional metadata to store with the request for tracing/distillation. reasoning_effort: anyOf: - type: string enum: - low - medium - high - xhigh - max - none - adaptive - type: integer - type: boolean - type: "null" title: Reasoning Effort description: >- Controls reasoning behavior for supported models. When enabled, the model's reasoning appears in the `reasoning_content` field of the response, separate from the final answer in `content`. **Accepted values:** - **String** (OpenAI-compatible): `'low'`, `'medium'`, `'high'`, or `'max'` to enable reasoning with varying effort levels; `'none'` to disable reasoning. - **Boolean** (Fireworks extension): `true` to enable reasoning, `false` to disable it. - **Integer** (Fireworks extension): A positive integer to set a hard token limit on reasoning output. Integer values enable the model's normal medium-style thinking behavior and force the model to end its thinking phase after at most that many generated thinking tokens. **Important:** Boolean values are normalized internally: `true` becomes `'medium'`, and `false` becomes `'none'`. This normalization happens before model-specific validation, so if a model doesn't support `'none'`, passing `false` will produce an error referencing `'none'`. **Model-specific behavior:** - **Qwen3**: Grammar-based reasoning on the reasoning-enabled `qwen3`/`qwen3p5` conversation styles. Older chat-mode Qwen3 deployments may opt into `qwen3-no-thinking`, which disables reasoning support. For reasoning-enabled styles, use `'none'` or `false` to disable. Supports integer token limits to cap reasoning output. `'low'`, `'medium'`, and `'high'` keep their model-specific behavior and are not hard budgets. - **MiniMax M2**: Reasoning is required (always on). Defaults to `'medium'` when omitted. Accepts only string `reasoning_effort`: `'low'`, `'medium'`, or `'high'`. `'none'` and boolean values are rejected. - **DeepSeek V3.1**: Binary on/off reasoning. Default reasoning off (matches chat template). Use `true`, `'low'`, `'medium'`, or `'high'` to enable; `'none'` or `false` to disable. - **DeepSeek V3.2**: Binary on/off reasoning. Default reasoning on. Use `'none'` or `false` to disable; effort levels and integers have no additional effect. - **DeepSeek V4**: Accepts `'none'`, `'low'`, `'medium'`, `'high'`, `'xhigh'`, and `'max'`. Default reasoning on (`'high'`). `'xhigh'` is silently promoted to `'max'`. `'max'` prepends a thorough-reasoning preamble; `'high'` enables thinking. `'low'` and `'medium'` are silently promoted to `'high'`. `'none'` or `false` disables thinking. - **GLM 4.5, GLM 4.5 Air, GLM 4.6, GLM 4.7, GLM 5.1**: Binary on/off reasoning. Default reasoning on. Use `'none'` or `false` to disable; effort levels and integers have no additional effect. - **GLM 5.2**: Two thinking tiers, `High` and `Max` (rendered as a `Reasoning Effort:` system line). `'high'` selects High; `'low'` and `'medium'` are collapsed to `'high'`; `'max'` and `'xhigh'` select Max; when omitted, the model default (`Max`) applies. `'none'` or `false` disables thinking. - **Harmony (OpenAI GPT-OSS 120B, GPT-OSS 20B)**: Accepts only `'low'`, `'medium'`, or `'high'`. Does not support `'none'`, `false`, or integer values — using these will return an error (e.g., "Invalid reasoning effort: none"). When omitted, defaults to `'medium'`. Lower effort produces faster responses with shorter reasoning. reasoning_history: anyOf: - type: string enum: - disabled - interleaved - preserved - type: "null" title: Reasoning History description: >- Controls how historical assistant reasoning content is included in the prompt for multi-turn conversations. **Accepted values:** - `null`: Use model/template default behavior (for **GLM-4.7**, the model/template default is `'interleaved'`, i.e. historical reasoning is cleared by default) - `'disabled'`: Strip `reasoning_content` from all messages before prompt construction - `'interleaved'`: Strip `reasoning_content` from messages up to (and including) the last user message - `'preserved'`: Preserve historical `reasoning_content` across the conversation **Model support:** | Model | Default | Supported values | | --- | --- | --- | | Kimi K2.7 | `'preserved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | Kimi K2.6 | `'interleaved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | Kimi K2 Instruct | `'preserved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | MiniMax M2 | `'interleaved'` | `'disabled'`, `'interleaved'` | | GLM-5.2 | `'interleaved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | GLM-4.7 | `'interleaved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | GLM-4.6 | `'interleaved'` | `'disabled'`, `'interleaved'` | | Qwen 3.6 | `'preserved'` | `'disabled'`, `'preserved'` | | DeepSeek V4 | `'interleaved'` | `'interleaved'` | For other models, refer to the model provider's documentation. **Note:** This parameter controls prompt formatting only. To disable reasoning computation entirely, use `reasoning_effort='none'`. thinking: anyOf: - $ref: "#/components/schemas/ThinkingConfigEnabled" - $ref: "#/components/schemas/ThinkingConfigDisabled" - $ref: "#/components/schemas/ThinkingConfigAdaptive" - type: "null" title: Thinking description: >- Configuration for enabling extended thinking (Anthropic-compatible format). This is an alternative to `reasoning_effort` for controlling reasoning behavior. **Format:** - `{"type": "enabled"}` - Enable thinking (equivalent to `reasoning_effort: true`) - `{"type": "enabled", "budget_tokens": }` - Enable thinking with a token budget (equivalent to `reasoning_effort: `). Must be >= 1024. - `{"type": "enabled", "keep": "all"}` - Enable thinking and preserve all historical reasoning content in the prompt (equivalent to `reasoning_history: "preserved"`). - `{"type": "disabled"}` - Disable thinking (equivalent to `reasoning_effort: "none"`) **Precedence with `reasoning_effort`:** `thinking.effort` (when set) overrides `reasoning_effort`; otherwise, for `type=enabled`, `reasoning_effort` is used as the effort level. `type=disabled` always disables thinking. return_token_ids: anyOf: - type: boolean - type: "null" title: Return Token Ids description: Return token IDs alongside text to avoid retokenization drift. default: false prompt_token_ids: anyOf: - items: type: integer type: array - type: "null" title: Prompt Token Ids description: >- Pre-tokenized prompt. When set, the server skips chat-template rendering and tokenization for the input — the supplied token IDs are fed directly to generation. Generation, the response formatter (incl. ``--formatter-backend vllm`` parsers), and tool-call routing all run as usual on the model's output. Intended for upstream gateways (e.g. dynamo) that own the tokenizer and chat template, and want this server only to generate and format. Mutually exclusive with ``messages``: passing both is a 400. functions: items: $ref: "#/components/schemas/ChatCompletionFunction" type: array title: Functions description: >- Deprecated in OpenAI. Use 'tools' instead. This will be automatically transformed to tools. deprecated: true prompt_truncate_len: anyOf: - type: integer - type: "null" title: Prompt Truncate Len description: >- The size (in tokens) to which to truncate chat prompts. This includes the system prompt (if any), previous user/assistant messages, and the current user message. Earlier user/assistant messages will be evicted first to fit the prompt into this length. The system prompt is preserved whenever possible and only truncated as a last resort. This should usually be set to a number much smaller << than the model's maximum context size, to allow enough remaining tokens for generating a response. If omitted, you may receive "prompt too long" errors in your responses as conversations grow. Note that even with this set, you may still receive "prompt too long" errors if individual messages (such as a very long system prompt or user message) exceed the model's context window on their own. parallel_tool_calls: anyOf: - type: boolean - type: "null" title: Parallel Tool Calls description: Enable parallel function calling. safe_tokenization: anyOf: - type: boolean - type: "null" title: Safe Tokenization description: >- When true, special tokens in user-provided content are never interpreted as actual special tokens during tokenization. This prevents prompt injection via special token strings (e.g. <|im_start|>, <｜User｜>). Supported for models using Jinja or HuggingFace chat templates with HuggingFace tokenizers. Explicit true returns an error if the model does not support it, or if combined with custom_chat_template on HuggingFace-backed models. Note: prompt_truncate_len is not applied when safe_tokenization is enabled. When omitted, safe tokenization is enabled by default on a best-effort basis: it is applied only where the model supports it and the request is text-only with no custom_chat_template and no prompt_truncate_len, and otherwise falls back to normal tokenization. function_call: anyOf: - type: string enum: - auto - none - $ref: "#/components/schemas/FunctionNameSpec" - type: "null" title: Function Call description: >- Deprecated in OpenAI. Use 'tool_choice' instead. This will be automatically transformed to tool_choice. deprecated: true additionalProperties: false type: object required: - model title: ChatCompletionRequest ChatCompletionTool: properties: type: type: string enum: - function title: Type description: The type of the tool. Currently, only `function` is supported. function: anyOf: - $ref: "#/components/schemas/ChatCompletionFunction" - type: "null" description: Required for function tools. additionalProperties: false type: object required: - type title: ChatCompletionTool ChatMessage: properties: role: type: string title: Role description: >- The role of the messages author. One of `system`, `user`, or `assistant`. content: anyOf: - type: string - items: $ref: "#/components/schemas/ChatMessageContent" type: array - type: "null" title: Content description: >- The contents of the message. `content` is required for all messages, and may be null for assistant messages with function calls. reasoning_content: anyOf: - type: string - type: "null" title: Reasoning Content description: >- The reasoning or thinking process generated by the model. This field is only available for certain reasoning models (GLM 4.5, GLM 4.5 Air, GPT OSS 120B, GPT OSS 20B) and contains the model's internal reasoning that would otherwise appear in `` tags within the content field. tool_calls: anyOf: - items: $ref: "#/components/schemas/ChatCompletionMessageToolCall" type: array - type: "null" title: Tool Calls description: The tool calls generated by the model, such as function calls. tool_call_id: anyOf: - type: string - type: "null" title: Tool Call Id additionalProperties: false type: object required: - role title: ChatMessage description: A chat completion message. ChatMessageContent: properties: type: type: string title: Type text: anyOf: - type: string - type: "null" title: Text image_url: anyOf: - $ref: "#/components/schemas/ChatMessageContentImageURL" - type: "null" video_url: anyOf: - $ref: "#/components/schemas/ChatMessageContentVideoURL" - type: "null" additionalProperties: false type: object required: - type title: ChatMessageContent ChatMessageContentImageURL: properties: url: type: string title: Url description: >- Image link or base64 data URI. ``mm_file://{file_id}`` is also accepted for assets uploaded via the Files API. detail: anyOf: - type: string - type: "null" title: Detail description: >- Detail level for image understanding. One of ``low`` / ``default`` / ``high`` (model-specific defaults table). Used to pick a default ``max_long_side_pixel`` when that field is absent. max_long_side_pixel: anyOf: - type: integer - type: "null" title: Max Long Side Pixel description: >- Per-image cap on the long side after resizing. When omitted, the model derives a default from ``detail``. Currently honored by MiniMax M3 VL preprocessing; other VL models ignore this field. See the M3 VL preprocessing spec (2026-05-29) for the full 3-step resize semantics (long-side cap → short-side floor at 112 px → hard total-pixel cap). additionalProperties: false type: object required: - url title: ChatMessageContentImageURL ChatMessageContentVideoURL: properties: url: type: string title: Url description: >- Video link or base64 data URI. ``mm_file://{file_id}`` is accepted for assets uploaded via the Files API (recommended for files > 50 MB). detail: anyOf: - type: string - type: "null" title: Detail description: >- Detail level for video understanding. One of ``low`` / ``default`` / ``high``. max_frames: anyOf: - type: integer - type: "null" title: Max Frames spatial_limit: anyOf: - type: integer - type: "null" title: Spatial Limit sample_fps: anyOf: - type: number - type: "null" title: Sample Fps description: >- Frame sampling rate (Kimi K2.5 VL legacy name). For MiniMax M3 VL prefer the canonical ``fps`` field. fps: anyOf: - type: number - type: "null" title: Fps description: >- Frame sampling rate, in [0.2, 5] Hz for MiniMax M3 VL (was [0.5, 2] in the pre-2026-05-29 spec). Higher values are more sensitive to motion at the cost of more tokens; lower values are cheaper but less responsive to fast scene changes. Equivalent to ``sample_fps`` on non-M3 video models. max_long_side_pixel: anyOf: - type: integer - type: "null" title: Max Long Side Pixel description: >- Per-frame cap on the long side after resizing. When omitted, the model derives a default from ``detail``. Currently honored by MiniMax M3 VL preprocessing; other VL models ignore this field. additionalProperties: false type: object required: - url title: ChatMessageContentVideoURL CompletionRequest: properties: model: type: string title: Model description: |- The name of the model to use. Example: `"accounts/fireworks/models/kimi-k2-instruct-0905"` user: anyOf: - type: string - type: "null" title: User description: >- A unique identifier representing your end-user, which can help monitor and detect abuse. prompt_cache_key: anyOf: - type: string - type: "null" title: Prompt Cache Key description: >- A key used for prompt caching session affinity. Requests with the same prompt_cache_key are routed to the same backend to maximize KV cache hit rates. This is the preferred field for session affinity (takes priority over the 'user' field). prompt_cache_isolation_key: anyOf: - type: string - type: "null" title: Prompt Cache Isolation Key description: Isolation key for prompt caching to separate cache entries. raw_output: anyOf: - type: boolean - type: "null" title: Raw Output description: Return raw output from the model. default: false perf_metrics_in_response: anyOf: - type: boolean - type: "null" title: Perf Metrics In Response description: >- Whether to include performance metrics in the response body. **Non-streaming requests:** Performance metrics are always included in response headers (e.g., `fireworks-prompt-tokens`, `fireworks-server-time-to-first-token`). Setting this to `true` additionally includes the same metrics in the response body under the `perf_metrics` field. **Streaming requests:** Performance metrics are only included in the response body under the `perf_metrics` field in the final chunk (when `finish_reason` is set). This is because headers may not be accessible during streaming. The response body `perf_metrics` field contains the following metrics: **Basic Metrics (all deployments):** - `prompt-tokens`: Number of tokens in the prompt - `cached-prompt-tokens`: Number of cached prompt tokens - `server-time-to-first-token`: Time from request start to first token (in seconds) - `server-processing-time`: Total processing time (in seconds, only for completed requests) **Predicted Outputs Metrics:** - `speculation-prompt-tokens`: Number of speculative prompt tokens - `speculation-prompt-matched-tokens`: Number of matched speculative prompt tokens (for completed requests) **Dedicated Deployment Only Metrics:** - `speculation-generated-tokens`: Number of speculative generated tokens (for completed requests) - `speculation-acceptance`: Speculation acceptance rates by position - `backend-host`: Hostname of the backend server - `num-concurrent-requests`: Number of concurrent requests - `deployment`: Deployment name - `tokenizer-queue-duration`: Time spent in tokenizer queue - `tokenizer-duration`: Time spent in tokenizer - `prefill-queue-duration`: Time spent in prefill queue - `prefill-duration`: Time spent in prefill - `generation-queue-duration`: Time spent in generation queue - `generation-duration`: Time spent in generation default: false stream: anyOf: - type: boolean - type: "null" title: Stream description: >- Whether to stream back partial progress. If set, tokens will be sent as data-only [server-sent events](https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format) as they become available, with the stream terminated by a `data: [DONE]` message. default: false stream_options: anyOf: - $ref: "#/components/schemas/StreamOptions" - type: "null" description: >- Options for streaming responses. Only valid with ``stream=true``. Fireworks includes a final SSE chunk carrying usage totals by default; set ``include_usage=false`` to opt out. The ``buffer_tokens`` / ``buffer_ms`` / ``buffer_mode`` fields coalesce SSE chunks by token count and/or time (vLLM backend; override the deployment default). "n": type: integer title: "N" description: >- How many completions to generate for each prompt. **Note:** Because this parameter generates many completions, it can quickly consume your token quota. Use carefully and ensure that you have reasonable settings for `max_tokens` and `stop`. Required range: `1 <= x <= 128` Example: `1` default: 1 service_tier: type: string enum: - auto - default - flex - priority title: Service Tier description: >- The service tier to use for the request. Specifies the processing type used for serving the request. Only "priority" is supported, while all other values will be treated as "default" tier. default: default stop: anyOf: - type: string - items: type: string type: array - type: "null" title: Stop description: >- Up to 4 sequences where the API will stop generating further tokens. The returned text will NOT contain the stop sequence. max_tokens: anyOf: - type: integer - type: "null" title: Max Tokens description: >- The maximum number of tokens to generate in the completion. If the token count of your prompt plus max_tokens exceeds the model's context length, the behavior depends on context_length_exceeded_behavior. By default, max_tokens will be lowered to fit in the context window instead of returning an error. max_completion_tokens: anyOf: - type: integer - type: "null" title: Max Completion Tokens description: Alias for max_tokens. Cannot be specified together with max_tokens. temperature: anyOf: - type: number - type: "null" title: Temperature description: >- What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. We generally recommend altering this or top_p but not both. Required range: `0 <= x <= 2` Example: `1` top_k: anyOf: - type: integer - type: "null" title: Top K description: >- Top-k sampling is another sampling method where the k most probable next tokens are filtered and the probability mass is redistributed among only those k next tokens. The value of k controls the number of candidates for the next token at each step during text generation. Must be between 0 and 100. Required range: `0 <= x <= 100` Example: `50` top_p: anyOf: - type: number - type: "null" title: Top P description: >- An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. We generally recommend altering this or temperature but not both. Required range: `0 <= x <= 1` Example: `1` min_p: anyOf: - type: number - type: "null" title: Min P description: >- Minimum probability threshold for token selection. Only tokens with probability >= min_p are considered for selection. This is an alternative to `top_p` and `top_k` sampling. Required range: `0 <= x <= 1` typical_p: anyOf: - type: number - type: "null" title: Typical P description: >- Typical-p sampling is an alternative to nucleus sampling. It considers the most typical tokens whose cumulative probability is at most typical_p. Required range: `0 <= x <= 1` frequency_penalty: anyOf: - type: number - type: "null" title: Frequency Penalty description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. Reasonable value is around 0.1 to 1 if the aim is to just reduce repetitive samples somewhat. If the aim is to strongly suppress repetition, then one can increase the coefficients up to 2, but this can noticeably degrade the quality of samples. Negative values can be used to increase the likelihood of repetition. See also `presence_penalty` for penalizing tokens that have at least one appearance at a fixed rate. OpenAI compatible (follows OpenAI's conventions for handling token frequency and repetition penalties). Required range: `-2 <= x <= 2` presence_penalty: anyOf: - type: number - type: "null" title: Presence Penalty description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Reasonable value is around 0.1 to 1 if the aim is to just reduce repetitive samples somewhat. If the aim is to strongly suppress repetition, then one can increase the coefficients up to 2, but this can noticeably degrade the quality of samples. Negative values can be used to increase the likelihood of repetition. See also `frequency_penalty` for penalizing tokens at an increasing rate depending on how often they appear. OpenAI compatible (follows OpenAI's conventions for handling token frequency and repetition penalties). Required range: `-2 <= x <= 2` repetition_penalty: anyOf: - type: number - type: "null" title: Repetition Penalty description: >- Applies a penalty to repeated tokens to discourage or encourage repetition. A value of `1.0` means no penalty, allowing free repetition. Values above `1.0` penalize repetition, reducing the likelihood of repeating tokens. Values between `0.0` and `1.0` reward repetition, increasing the chance of repeated tokens. For a good balance, a value of `1.2` is often recommended. Note that the penalty is applied to both the generated output and the prompt in decoder-only models. Required range: `0 <= x <= 2` mirostat_target: anyOf: - type: number - type: "null" title: Mirostat Target description: >- Defines the target perplexity for the Mirostat algorithm. Perplexity measures the unpredictability of the generated text, with higher values encouraging more diverse and creative outputs, while lower values prioritize predictability and coherence. The algorithm dynamically adjusts the token selection to maintain this target during text generation. If not specified, Mirostat sampling is disabled. mirostat_lr: anyOf: - type: number - type: "null" title: Mirostat Lr description: >- Specifies the learning rate for the Mirostat sampling algorithm, which controls how quickly the model adjusts its token distribution to maintain the target perplexity. A smaller value slows down the adjustments, leading to more stable but gradual shifts, while higher values speed up corrections at the cost of potential instability. seed: anyOf: - type: integer - type: "null" title: Seed description: Random seed for deterministic sampling. logprobs: anyOf: - type: integer - type: boolean - type: "null" title: Logprobs description: >- Include log probabilities in the response. This accepts either a boolean or an integer: If set to `true`, log probabilities are included and the number of alternatives can be controlled via `top_logprobs` (OpenAI-compatible behavior). If set to an integer N, include log probabilities for up to N most likely tokens per position in the legacy format. N must be between 0 and the deployment's `--max-logprobs` limit (5 by default). The API will always return the logprob of the sampled token, so there may be up to `logprobs+1` elements in the response when an integer is used. top_logprobs: anyOf: - type: integer - type: "null" title: Top Logprobs description: >- An integer specifying the number of most likely tokens to return at each token position, each with an associated log probability. Must be between 0 and the deployment's `--max-logprobs` limit (5 by default). When `logprobs` is set, `top_logprobs` can be used to modify how many top log probabilities are returned. If `top_logprobs` is not set, the API will return up to `logprobs` tokens per position. sampling_mask: anyOf: - type: string enum: - count - non_zero_list - non_zero_buffer - type: "null" title: Sampling Mask description: >- Opt-in sampling mask metadata for generated tokens. When set to `"count"`, each generated token in the new logprobs format includes the number of token logits still eligible for sampling after filters such as top_p and top_k are applied. `"non_zero_list"` additionally returns active token IDs in `sampling_mask`; `"non_zero_buffer"` additionally returns a base64-encoded little-endian uint32 buffer of active token IDs. Non-zero payloads are omitted for positions with more active tokens than 1000. echo: anyOf: - type: boolean - type: "null" title: Echo description: Echo back the prompt in addition to the completion. default: false echo_last: anyOf: - type: integer - type: "null" title: Echo Last description: >- Echo back the last N tokens of the prompt in addition to the completion. This is useful for obtaining logprobs of the prompt suffix but without transferring too much data. Passing `echo_last=len(prompt)` is the same as `echo=True` ignore_eos: type: boolean title: Ignore Eos description: >- This setting controls whether the model should ignore the End of Sequence (EOS) token. When set to `True`, the model will continue generating tokens even after the EOS token is produced. By default, it stops when the EOS token is reached. default: false context_length_exceeded_behavior: type: string enum: - error - truncate title: Context Length Exceeded Behavior description: >- What to do if the token count of prompt plus `max_tokens` exceeds the model's context window. Passing `truncate` limits the `max_tokens` to at most `context_window_length - prompt_length`. This is the default. Passing `error` would trigger a request error. The default of `'truncate'` is selected as it allows to ask for high `max_tokens` value while respecting the context window length without having to do client-side prompt tokenization. Note, that it differs from OpenAI's behavior that matches that of `error`. default: truncate response_format: anyOf: - $ref: "#/components/schemas/ResponseFormat" - type: "null" description: >- Allows to force the model to produce specific output format. Setting to `{ "type": "json_object" }` enables JSON mode, which guarantees the message the model generates is valid JSON. If `"type"` is `"json_schema"`, a JSON schema must be provided. E.g., `response_format = {"type": "json_schema", "json_schema": }`. Important: when using JSON mode, it's crucial to also instruct the model to produce JSON via a system or user message. Without this, the model may generate an unending stream of whitespace until the generation reaches the token limit, resulting in a long-running and seemingly "stuck" request. Also note that the message content may be partially cut off if `finish_reason="length"`, which indicates the generation exceeded `max_tokens` or the conversation exceeded the max context length. In this case the return value might not be a valid JSON. logit_bias: anyOf: - additionalProperties: type: number type: object - type: "null" title: Logit Bias description: >- Modify the likelihood of specified tokens appearing in the completion. Accepts a json object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. speculation: anyOf: - type: string - items: type: integer type: array - type: "null" title: Speculation description: Speculative decoding prompt or token IDs to speed up generation. prediction: anyOf: - $ref: "#/components/schemas/PredictedOutput" - type: string - type: "null" title: Prediction description: >- OpenAI-compatible predicted output for speculative decoding. Can be a PredictedOutput object or a simple string. Automatically transformed to speculation. metadata: anyOf: - additionalProperties: type: string type: object - type: "null" title: Metadata description: >- Additional metadata to store with the request for tracing/distillation. reasoning_effort: anyOf: - type: string enum: - low - medium - high - xhigh - max - none - adaptive - type: integer - type: boolean - type: "null" title: Reasoning Effort description: >- Controls reasoning behavior for supported models. When enabled, the model's reasoning appears in the `reasoning_content` field of the response, separate from the final answer in `content`. **Accepted values:** - **String** (OpenAI-compatible): `'low'`, `'medium'`, `'high'`, or `'max'` to enable reasoning with varying effort levels; `'none'` to disable reasoning. - **Boolean** (Fireworks extension): `true` to enable reasoning, `false` to disable it. - **Integer** (Fireworks extension): A positive integer to set a hard token limit on reasoning output. Integer values enable the model's normal medium-style thinking behavior and force the model to end its thinking phase after at most that many generated thinking tokens. **Important:** Boolean values are normalized internally: `true` becomes `'medium'`, and `false` becomes `'none'`. This normalization happens before model-specific validation, so if a model doesn't support `'none'`, passing `false` will produce an error referencing `'none'`. **Model-specific behavior:** - **Qwen3**: Grammar-based reasoning on the reasoning-enabled `qwen3`/`qwen3p5` conversation styles. Older chat-mode Qwen3 deployments may opt into `qwen3-no-thinking`, which disables reasoning support. For reasoning-enabled styles, use `'none'` or `false` to disable. Supports integer token limits to cap reasoning output. `'low'`, `'medium'`, and `'high'` keep their model-specific behavior and are not hard budgets. - **MiniMax M2**: Reasoning is required (always on). Defaults to `'medium'` when omitted. Accepts only string `reasoning_effort`: `'low'`, `'medium'`, or `'high'`. `'none'` and boolean values are rejected. - **DeepSeek V3.1**: Binary on/off reasoning. Default reasoning off (matches chat template). Use `true`, `'low'`, `'medium'`, or `'high'` to enable; `'none'` or `false` to disable. - **DeepSeek V3.2**: Binary on/off reasoning. Default reasoning on. Use `'none'` or `false` to disable; effort levels and integers have no additional effect. - **DeepSeek V4**: Accepts `'none'`, `'low'`, `'medium'`, `'high'`, `'xhigh'`, and `'max'`. Default reasoning on (`'high'`). `'xhigh'` is silently promoted to `'max'`. `'max'` prepends a thorough-reasoning preamble; `'high'` enables thinking. `'low'` and `'medium'` are silently promoted to `'high'`. `'none'` or `false` disables thinking. - **GLM 4.5, GLM 4.5 Air, GLM 4.6, GLM 4.7, GLM 5.1**: Binary on/off reasoning. Default reasoning on. Use `'none'` or `false` to disable; effort levels and integers have no additional effect. - **GLM 5.2**: Two thinking tiers, `High` and `Max` (rendered as a `Reasoning Effort:` system line). `'high'` selects High; `'low'` and `'medium'` are collapsed to `'high'`; `'max'` and `'xhigh'` select Max; when omitted, the model default (`Max`) applies. `'none'` or `false` disables thinking. - **Harmony (OpenAI GPT-OSS 120B, GPT-OSS 20B)**: Accepts only `'low'`, `'medium'`, or `'high'`. Does not support `'none'`, `false`, or integer values — using these will return an error (e.g., "Invalid reasoning effort: none"). When omitted, defaults to `'medium'`. Lower effort produces faster responses with shorter reasoning. reasoning_history: anyOf: - type: string enum: - disabled - interleaved - preserved - type: "null" title: Reasoning History description: >- Controls how historical assistant reasoning content is included in the prompt for multi-turn conversations. **Accepted values:** - `null`: Use model/template default behavior (for **GLM-4.7**, the model/template default is `'interleaved'`, i.e. historical reasoning is cleared by default) - `'disabled'`: Strip `reasoning_content` from all messages before prompt construction - `'interleaved'`: Strip `reasoning_content` from messages up to (and including) the last user message - `'preserved'`: Preserve historical `reasoning_content` across the conversation **Model support:** | Model | Default | Supported values | | --- | --- | --- | | Kimi K2.7 | `'preserved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | Kimi K2.6 | `'interleaved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | Kimi K2 Instruct | `'preserved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | MiniMax M2 | `'interleaved'` | `'disabled'`, `'interleaved'` | | GLM-5.2 | `'interleaved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | GLM-4.7 | `'interleaved'` | `'disabled'`, `'interleaved'`, `'preserved'` | | GLM-4.6 | `'interleaved'` | `'disabled'`, `'interleaved'` | | Qwen 3.6 | `'preserved'` | `'disabled'`, `'preserved'` | | DeepSeek V4 | `'interleaved'` | `'interleaved'` | For other models, refer to the model provider's documentation. **Note:** This parameter controls prompt formatting only. To disable reasoning computation entirely, use `reasoning_effort='none'`. thinking: anyOf: - $ref: "#/components/schemas/ThinkingConfigEnabled" - $ref: "#/components/schemas/ThinkingConfigDisabled" - $ref: "#/components/schemas/ThinkingConfigAdaptive" - type: "null" title: Thinking description: >- Configuration for enabling extended thinking (Anthropic-compatible format). This is an alternative to `reasoning_effort` for controlling reasoning behavior. **Format:** - `{"type": "enabled"}` - Enable thinking (equivalent to `reasoning_effort: true`) - `{"type": "enabled", "budget_tokens": }` - Enable thinking with a token budget (equivalent to `reasoning_effort: `). Must be >= 1024. - `{"type": "enabled", "keep": "all"}` - Enable thinking and preserve all historical reasoning content in the prompt (equivalent to `reasoning_history: "preserved"`). - `{"type": "disabled"}` - Disable thinking (equivalent to `reasoning_effort: "none"`) **Precedence with `reasoning_effort`:** `thinking.effort` (when set) overrides `reasoning_effort`; otherwise, for `type=enabled`, `reasoning_effort` is used as the effort level. `type=disabled` always disables thinking. return_token_ids: anyOf: - type: boolean - type: "null" title: Return Token Ids description: Return token IDs alongside text to avoid retokenization drift. default: false prompt: anyOf: - type: string - items: type: string type: array - items: type: integer type: array - items: items: type: integer type: array type: array title: Prompt description: >- The prompt to generate completions for. It can be a single string or an array of strings. It can also be an array of integers or an array of integer arrays, which allows to pass already tokenized prompt. If multiple prompts are specified, several choices with corresponding `index` will be returned in the output. images: anyOf: - items: type: string type: array - items: items: type: string type: array type: array - type: "null" title: Images description: >- The list of base64 encoded images for visual language completition generation. They should be formatted as MIME_TYPE, eg. data:image/jpeg;base64, Additionally, the number of images provided should match the number of image placeholder tokens in the prompt (string prompts: '' or model-specific pads such as '<|image_pad|>'; tokenized prompts: one image pad token ID per image, unexpanded). additionalProperties: false type: object required: - prompt - model title: CompletionRequest EmbeddingRequest: properties: model: type: string title: Model description: The name of the model to use. user: anyOf: - type: string - type: "null" title: User description: >- A unique identifier representing your end-user, which can help monitor and detect abuse. prompt_cache_key: anyOf: - type: string - type: "null" title: Prompt Cache Key description: >- A key used for prompt caching session affinity. Requests with the same prompt_cache_key are routed to the same backend to maximize KV cache hit rates. This is the preferred field for session affinity (takes priority over the 'user' field). prompt_cache_isolation_key: anyOf: - type: string - type: "null" title: Prompt Cache Isolation Key description: Isolation key for prompt caching to separate cache entries. raw_output: anyOf: - type: boolean - type: "null" title: Raw Output description: Return raw output from the model. default: false perf_metrics_in_response: anyOf: - type: boolean - type: "null" title: Perf Metrics In Response description: >- Whether to include performance metrics in the response body. **Non-streaming requests:** Performance metrics are always included in response headers (e.g., `fireworks-prompt-tokens`, `fireworks-server-time-to-first-token`). Setting this to `true` additionally includes the same metrics in the response body under the `perf_metrics` field. **Streaming requests:** Performance metrics are only included in the response body under the `perf_metrics` field in the final chunk (when `finish_reason` is set). This is because headers may not be accessible during streaming. The response body `perf_metrics` field contains the following metrics: **Basic Metrics (all deployments):** - `prompt-tokens`: Number of tokens in the prompt - `cached-prompt-tokens`: Number of cached prompt tokens - `server-time-to-first-token`: Time from request start to first token (in seconds) - `server-processing-time`: Total processing time (in seconds, only for completed requests) **Predicted Outputs Metrics:** - `speculation-prompt-tokens`: Number of speculative prompt tokens - `speculation-prompt-matched-tokens`: Number of matched speculative prompt tokens (for completed requests) **Dedicated Deployment Only Metrics:** - `speculation-generated-tokens`: Number of speculative generated tokens (for completed requests) - `speculation-acceptance`: Speculation acceptance rates by position - `backend-host`: Hostname of the backend server - `num-concurrent-requests`: Number of concurrent requests - `deployment`: Deployment name - `tokenizer-queue-duration`: Time spent in tokenizer queue - `tokenizer-duration`: Time spent in tokenizer - `prefill-queue-duration`: Time spent in prefill queue - `prefill-duration`: Time spent in prefill - `generation-queue-duration`: Time spent in generation queue - `generation-duration`: Time spent in generation default: false input: anyOf: - type: string - items: type: string type: array - items: items: type: integer type: array type: array - items: type: integer type: array - additionalProperties: true type: object - items: additionalProperties: true type: object type: array title: Input prompt_template: anyOf: - type: string - type: "null" title: Prompt Template encoding_format: type: string enum: - float - base64 title: Encoding Format default: float dimensions: anyOf: - type: integer - type: "null" title: Dimensions return_logits: anyOf: - items: type: integer type: array - type: "null" title: Return Logits normalize: anyOf: - type: boolean - type: "null" title: Normalize default: false fanout_size: anyOf: - type: integer - type: "null" title: Fanout Size additionalProperties: false type: object required: - input - model title: EmbeddingRequest FunctionNameSpec: properties: name: type: string title: Name additionalProperties: false type: object required: - name title: FunctionNameSpec FunctionSelection: properties: type: type: string const: function title: Type function: anyOf: - $ref: "#/components/schemas/FunctionNameSpec" - type: "null" additionalProperties: false type: object required: - type title: FunctionSelection PredictedOutput: properties: content: anyOf: - type: string - items: $ref: "#/components/schemas/ChatMessageContent" type: array title: Content type: type: string const: content title: Type default: content additionalProperties: false type: object required: - content title: PredictedOutput description: OpenAI-compatible struct for the "speculation" field. RerankRequestBody: properties: model: anyOf: - type: string - type: "null" title: Model query: type: string title: Query documents: items: type: string type: array title: Documents top_n: anyOf: - type: integer - type: "null" title: Top N return_documents: type: boolean title: Return Documents default: true task: anyOf: - type: string - type: "null" title: Task type: object required: - query - documents title: RerankRequestBody ResponseFormat: properties: type: type: string enum: - json_object - json_schema - grammar - text title: Type schema: anyOf: - additionalProperties: true type: object - type: string - type: "null" title: Schema grammar: anyOf: - type: string - type: "null" title: Grammar json_schema: anyOf: - additionalProperties: true type: object - type: string - type: "null" title: Json Schema additionalProperties: false type: object required: - type title: ResponseFormat StreamOptions: properties: include_usage: anyOf: - type: boolean - type: "null" title: Include Usage description: >- Whether to include a trailing SSE chunk with usage totals (with an empty `choices` array). Unlike the OpenAI spec, Fireworks includes usage by default for streaming responses; set this to `false` to opt out. When emitted, usage rides a separate final chunk (before `data: [DONE]`), not the chunk carrying `finish_reason`. include_internal_content: anyOf: - type: boolean - type: "null" title: Include Internal Content description: >- When true, include an `internal_content` object (currently `token_id`) inside each streaming delta — equivalent to `return_token_ids`, emitted under `choices[].delta.internal_content`. Omitted entirely from response chunks when false. default: false buffer_tokens: anyOf: - type: integer minimum: 0 - type: "null" title: Buffer Tokens description: >- Coalesce streaming SSE chunks until this many text deltas (~tokens) accumulate before flushing a merged chunk. 0 disables the token threshold. Honored only on the vLLM backend; overrides the deployment default when set. buffer_ms: anyOf: - type: number minimum: 0 - type: "null" title: Buffer Ms description: >- Coalesce streaming SSE chunks for up to this many milliseconds before flushing a merged chunk. 0 disables the time threshold. Honored only on the vLLM backend; overrides the deployment default when set. buffer_mode: anyOf: - type: string enum: - any - all - type: "null" title: Buffer Mode description: >- When both buffer_tokens and buffer_ms are set: 'any' flushes when either threshold is reached; 'all' flushes only when both are. Overrides the deployment default when set; defaults to 'any'. additionalProperties: false type: object title: StreamOptions TgiGenerateParameters: properties: details: type: boolean title: Details default: false max_new_tokens: anyOf: - type: integer - type: "null" title: Max New Tokens repetition_penalty: anyOf: - type: number - type: "null" title: Repetition Penalty return_full_text: type: boolean title: Return Full Text default: false stop: anyOf: - items: type: string type: array - type: "null" title: Stop temperature: anyOf: - type: number - type: "null" title: Temperature top_k: anyOf: - type: integer - type: "null" title: Top K top_p: anyOf: - type: number - type: "null" title: Top P truncate: anyOf: - type: integer - type: "null" title: Truncate typical_p: anyOf: - type: number - type: "null" title: Typical P additionalProperties: false type: object title: TgiGenerateParameters TgiGenerateRequest: properties: model: type: string title: Model description: The name of the model to use. user: anyOf: - type: string - type: "null" title: User description: >- A unique identifier representing your end-user, which can help monitor and detect abuse. prompt_cache_key: anyOf: - type: string - type: "null" title: Prompt Cache Key description: >- A key used for prompt caching session affinity. Requests with the same prompt_cache_key are routed to the same backend to maximize KV cache hit rates. This is the preferred field for session affinity (takes priority over the 'user' field). prompt_cache_isolation_key: anyOf: - type: string - type: "null" title: Prompt Cache Isolation Key description: Isolation key for prompt caching to separate cache entries. raw_output: anyOf: - type: boolean - type: "null" title: Raw Output description: Return raw output from the model. default: false perf_metrics_in_response: anyOf: - type: boolean - type: "null" title: Perf Metrics In Response description: >- Whether to include performance metrics in the response body. **Non-streaming requests:** Performance metrics are always included in response headers (e.g., `fireworks-prompt-tokens`, `fireworks-server-time-to-first-token`). Setting this to `true` additionally includes the same metrics in the response body under the `perf_metrics` field. **Streaming requests:** Performance metrics are only included in the response body under the `perf_metrics` field in the final chunk (when `finish_reason` is set). This is because headers may not be accessible during streaming. The response body `perf_metrics` field contains the following metrics: **Basic Metrics (all deployments):** - `prompt-tokens`: Number of tokens in the prompt - `cached-prompt-tokens`: Number of cached prompt tokens - `server-time-to-first-token`: Time from request start to first token (in seconds) - `server-processing-time`: Total processing time (in seconds, only for completed requests) **Predicted Outputs Metrics:** - `speculation-prompt-tokens`: Number of speculative prompt tokens - `speculation-prompt-matched-tokens`: Number of matched speculative prompt tokens (for completed requests) **Dedicated Deployment Only Metrics:** - `speculation-generated-tokens`: Number of speculative generated tokens (for completed requests) - `speculation-acceptance`: Speculation acceptance rates by position - `backend-host`: Hostname of the backend server - `num-concurrent-requests`: Number of concurrent requests - `deployment`: Deployment name - `tokenizer-queue-duration`: Time spent in tokenizer queue - `tokenizer-duration`: Time spent in tokenizer - `prefill-queue-duration`: Time spent in prefill queue - `prefill-duration`: Time spent in prefill - `generation-queue-duration`: Time spent in generation queue - `generation-duration`: Time spent in generation default: false inputs: type: string title: Inputs parameters: anyOf: - $ref: "#/components/schemas/TgiGenerateParameters" - type: "null" stream: anyOf: - type: boolean - type: "null" title: Stream default: false additionalProperties: false type: object required: - inputs - model title: TgiGenerateRequest ThinkingConfigAdaptive: properties: type: type: string const: adaptive title: Type additionalProperties: false type: object required: - type title: ThinkingConfigAdaptive description: >- Configuration that lets the model decide whether to think (MiniMax M3). Per the M3 API spec (§1.4), `{"type": "adaptive"}` is M3's default — the model decides whether to emit a thinking phase. No forced first token is applied. Currently accepted only by MiniMax M3; other model families reject it. ThinkingConfigDisabled: properties: type: type: string const: disabled title: Type additionalProperties: false type: object required: - type title: ThinkingConfigDisabled description: >- Configuration for disabling extended thinking (Anthropic-compatible format). ThinkingConfigEnabled: properties: type: type: string const: enabled title: Type default: enabled budget_tokens: anyOf: - type: integer - type: "null" title: Budget Tokens description: >- Determines how many tokens the model can use for its internal reasoning process. Larger budgets can enable more thorough analysis for complex problems, improving response quality. Must be >= 1024 if specified. keep: anyOf: - type: string const: all - type: "null" title: Keep description: >- Controls whether historical reasoning content is preserved in the prompt. When set to `"all"`, all previous assistant turns' reasoning_content is included in the rendered prompt (equivalent to `reasoning_history: "preserved"`). When omitted (null), the model uses its default history behavior. Only valid when `type` is `"enabled"`. budget_end_str: anyOf: - type: string - type: "null" title: Budget End Str description: >- Natural-language transition phrase that the model is forced to emit just before the end-thinking token (``) when `budget_tokens` is exhausted. This produces a more natural conclusion than a hard token slam (matches vLLM's `reasoning_end_str` behavior). Defaults to a built-in phrase. Set to "" to disable the bridge and force `` immediately. Only meaningful when `budget_tokens` is set. effort: anyOf: - type: string enum: - low - medium - high - xhigh - max - none - adaptive - type: integer - type: boolean - type: "null" title: Effort description: >- Reasoning effort level (Kimi/Moonshot spec). Accepts the same values as top-level `reasoning_effort`; when set it takes precedence over (overwrites) `reasoning_effort`. additionalProperties: false type: object title: ThinkingConfigEnabled description: >- Configuration for enabling extended thinking (Anthropic-compatible format). CompletionResponse: additionalProperties: false description: The response message from a /v1/completions call. properties: id: description: A unique identifier of the response title: Id type: string object: default: text_completion description: The object type, which is always "text_completion" title: Object type: string created: description: The Unix time in seconds when the response was generated title: Created type: integer model: description: The model used for the completion title: Model type: string choices: description: The list of generated completion choices items: $ref: "#/components/schemas/Choice" title: Choices type: array usage: $ref: "#/components/schemas/UsageInfo" description: Usage statistics for the completion perf_metrics: anyOf: - additionalProperties: true type: object - type: "null" default: null description: >- See parameter [perf_metrics_in_response](#body-perf-metrics-in-response) title: Perf Metrics required: - id - created - model - choices - usage title: CompletionResponse type: object Choice: additionalProperties: false description: A completion choice. properties: index: description: The index of the completion choice title: Index type: integer text: description: The completion response title: Text type: string logprobs: anyOf: - $ref: "#/components/schemas/LogProbs" - $ref: "#/components/schemas/NewLogProbs" - type: "null" default: null description: The log probabilities of the most likely tokens title: Logprobs finish_reason: anyOf: - enum: - stop - length - error type: string - type: "null" default: null description: >- The reason the model stopped generating tokens. This will be "stop" if the model hit a natural stop point or a provided stop sequence, or "length" if the maximum number of tokens specified in the request was reached title: Finish Reason raw_output: anyOf: - $ref: "#/components/schemas/RawOutput" - type: "null" default: null prompt_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for the prompt (when return_token_ids=true) title: Prompt Token Ids token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for the generated completion (when return_token_ids=true) title: Token Ids required: - index - text title: Choice type: object LogProbs: additionalProperties: false description: Legacy log probabilities format properties: tokens: items: type: string title: Tokens type: array token_logprobs: items: type: number title: Token Logprobs type: array top_logprobs: anyOf: - items: additionalProperties: type: number type: object type: array - type: "null" title: Top Logprobs text_offset: items: type: integer title: Text Offset type: array token_ids: anyOf: - items: type: integer type: array - type: "null" default: null title: Token Ids title: LogProbs type: object NewLogProbs: additionalProperties: false description: OpenAI-compatible log probabilities format properties: content: items: $ref: "#/components/schemas/NewLogProbsContent" title: Content type: array title: NewLogProbs type: object NewLogProbsContent: additionalProperties: false properties: token: title: Token type: string logprob: title: Logprob type: number sampling_logprob: anyOf: - type: number - type: "null" title: Sampling Logprob sampling_mask_count: anyOf: - type: integer - type: "null" default: null title: Sampling Mask Count sampling_mask: anyOf: - items: type: integer type: array - type: string - type: "null" default: null title: Sampling Mask bytes: items: type: integer title: Bytes type: array top_logprobs: items: $ref: "#/components/schemas/NewLogProbsContentTopLogProbs" title: Top Logprobs type: array token_id: title: Token Id type: integer text_offset: title: Text Offset type: integer last_activation: anyOf: - type: string - type: "null" default: null title: Last Activation routing_matrix: anyOf: - type: string - type: "null" default: null title: Routing Matrix required: - token - logprob - sampling_logprob - bytes - token_id - text_offset title: NewLogProbsContent type: object NewLogProbsContentTopLogProbs: additionalProperties: false properties: token: title: Token type: string logprob: title: Logprob type: number token_id: title: Token Id type: integer bytes: items: type: integer title: Bytes type: array required: - token - logprob - token_id title: NewLogProbsContentTopLogProbs type: object PromptTokensDetails: additionalProperties: false properties: cached_tokens: anyOf: - type: integer - type: "null" default: null title: Cached Tokens title: PromptTokensDetails type: object RawOutput: additionalProperties: false description: |- Extension of OpenAI that returns low-level interaction of what the model sees, including the formatted prompt and function calls properties: prompt_fragments: description: >- Pieces of the prompt (like individual messages) before truncation and concatenation. Depending on prompt_truncate_len some of the messages might be dropped. Contains a mix of strings to be tokenized and individual tokens (if dictated by the conversation template) items: anyOf: - type: string - type: integer title: Prompt Fragments type: array prompt_token_ids: description: Fully processed prompt as seen by the model items: type: integer title: Prompt Token Ids type: array completion: description: >- Raw completion produced by the model before any tool calls are parsed title: Completion type: string completion_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for the raw completion title: Completion Token Ids completion_logprobs: anyOf: - $ref: "#/components/schemas/NewLogProbs" - type: "null" default: null description: >- Log probabilities for the completion. Only populated if logprobs is specified in the request images: anyOf: - items: type: string type: array - type: "null" default: null description: Images in the prompt title: Images videos: anyOf: - items: type: string type: array - type: "null" default: null description: Videos in the prompt title: Videos grammar: anyOf: - type: string - type: "null" default: null description: >- Grammar used for constrained decoding, can be either user provided (directly or JSON schema) or inferred by the chat template title: Grammar required: - prompt_fragments - prompt_token_ids - completion title: RawOutput type: object UsageInfo: additionalProperties: false description: Usage statistics. properties: prompt_tokens: description: The number of tokens in the prompt title: Prompt Tokens type: integer total_tokens: description: The total number of tokens used in the request (prompt + completion) title: Total Tokens type: integer completion_tokens: anyOf: - type: integer - type: "null" default: null description: The number of tokens in the generated completion title: Completion Tokens prompt_tokens_details: anyOf: - $ref: "#/components/schemas/PromptTokensDetails" - type: "null" default: null description: Details about prompt tokens, including cached tokens required: - prompt_tokens - total_tokens title: UsageInfo type: object CompletionStreamResponse: additionalProperties: false description: The streamed response message from a /v1/completions call. properties: id: description: A unique identifier of the response title: Id type: string object: default: text_completion description: The object type, which is always "text_completion" title: Object type: string created: description: The Unix time in seconds when the response was generated title: Created type: integer model: description: The model used for the chat completion title: Model type: string choices: description: The list of streamed completion choices items: $ref: "#/components/schemas/CompletionResponseStreamChoice" title: Choices type: array usage: anyOf: - $ref: "#/components/schemas/UsageInfo" - type: "null" default: null perf_metrics: anyOf: - additionalProperties: true type: object - type: "null" default: null description: >- See parameter [perf_metrics_in_response](#body-perf-metrics-in-response) title: Perf Metrics required: - id - created - model - choices title: CompletionStreamResponse type: object CompletionResponseStreamChoice: additionalProperties: false description: |- A streamed completion choice. Attributes: index (int): The index of the completion choice. text (str): The completion response. logprobs (float, optional): The log probabilities of the most likely tokens. finish_reason (str): The reason the model stopped generating tokens. This will be "stop" if the model hit a natural stop point or a provided stop sequence, or "length" if the maximum number of tokens specified in the request was reached. prompt_token_ids (Optional[List[int]]): Token IDs for the prompt (when return_token_ids=true, sent in first chunk) token_ids (Optional[List[int]]): Token IDs for this chunk (when return_token_ids=true) properties: index: title: Index type: integer text: title: Text type: string logprobs: anyOf: - $ref: "#/components/schemas/LogProbs" - $ref: "#/components/schemas/NewLogProbs" - type: "null" default: null title: Logprobs finish_reason: anyOf: - enum: - stop - length - error type: string - type: "null" default: null title: Finish Reason raw_output: anyOf: - $ref: "#/components/schemas/RawOutput" - type: "null" default: null prompt_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null title: Prompt Token Ids token_ids: anyOf: - items: type: integer type: array - type: "null" default: null title: Token Ids required: - index - text title: CompletionResponseStreamChoice type: object ChatCompletionResponse: additionalProperties: false description: The response message from a /v1/chat/completions call. properties: id: description: A unique identifier of the response title: Id type: string object: default: chat.completion description: The object type, which is always "chat.completion" title: Object type: string created: description: The Unix time in seconds when the response was generated title: Created type: integer model: description: The model used for the chat completion title: Model type: string choices: description: The list of chat completion choices items: $ref: "#/components/schemas/ChatCompletionResponseChoice" title: Choices type: array usage: anyOf: - $ref: "#/components/schemas/UsageInfo" - type: "null" default: null perf_metrics: anyOf: - additionalProperties: true type: object - type: "null" default: null description: >- See parameter [perf_metrics_in_response](#body-perf-metrics-in-response) title: Perf Metrics prompt_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for the prompt (when return_token_ids=true) title: Prompt Token Ids required: - id - created - model - choices title: ChatCompletionResponse type: object ChatCompletionResponseChoice: additionalProperties: false description: |- A chat completion choice generated by a chat model. Attributes: index (int): The index of the chat completion choice. message (ChatMessage): The chat completion message. finish_reason (Optional[str]): The reason the model stopped generating tokens. This will be "stop" if the model hit a natural stop point or a provided stop sequence, or "length" if the maximum number of tokens specified in the request was reached. token_ids (Optional[List[int]]): Token IDs for the generated message content (when return_token_ids=true) properties: index: title: Index type: integer message: $ref: "#/components/schemas/ChatMessage" finish_reason: anyOf: - type: string - type: "null" default: null title: Finish Reason logprobs: anyOf: - $ref: "#/components/schemas/LogProbs" - $ref: "#/components/schemas/NewLogProbs" - type: "null" default: null title: Logprobs raw_output: anyOf: - $ref: "#/components/schemas/RawOutput" - type: "null" default: null token_ids: anyOf: - items: type: integer type: array - type: "null" default: null title: Token Ids required: - index - message title: ChatCompletionResponseChoice type: object ChatCompletionStreamResponse: additionalProperties: false description: The streamed response message from a /v1/chat/completions call. properties: id: description: A unique identifier of the response title: Id type: string object: default: chat.completion.chunk description: The object type, which is always "chat.completion.chunk" title: Object type: string created: description: The Unix time in seconds when the response was generated title: Created type: integer model: description: The model used for the chat completion title: Model type: string choices: description: The list of streamed chat completion choices items: $ref: "#/components/schemas/ChatCompletionResponseStreamChoice" title: Choices type: array usage: anyOf: - $ref: "#/components/schemas/UsageInfo" - type: "null" default: null perf_metrics: anyOf: - additionalProperties: true type: object - type: "null" default: null description: >- See parameter [perf_metrics_in_response](#body-perf-metrics-in-response) title: Perf Metrics prompt_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: >- Token IDs for the prompt (when return_token_ids=true, sent in first chunk) title: Prompt Token Ids required: - id - created - model - choices title: ChatCompletionStreamResponse type: object ChatCompletionResponseStreamChoice: additionalProperties: false description: A streamed chat completion choice. properties: index: description: The index of the chat completion choice title: Index type: integer delta: $ref: "#/components/schemas/DeltaMessage" description: The message delta finish_reason: anyOf: - enum: - stop - length - function_call - tool_calls type: string - type: "null" default: null description: >- The reason the model stopped generating tokens. This will be "stop" if the model hit a natural stop point or a provided stop sequence, or "length" if the maximum number of tokens specified in the request was reached title: Finish Reason logprobs: anyOf: - $ref: "#/components/schemas/LogProbs" - $ref: "#/components/schemas/NewLogProbs" - type: "null" default: null title: Logprobs raw_output: anyOf: - $ref: "#/components/schemas/RawOutput" - type: "null" default: null prompt_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for the prompt (when return_token_ids=true) title: Prompt Token Ids token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for this chunk (when return_token_ids=true) title: Token Ids usage: anyOf: - $ref: "#/components/schemas/UsageInfo" - type: "null" default: null description: >- Usage for this choice, emitted on its final (finish) chunk (deployment-gated extension). required: - index - delta title: ChatCompletionResponseStreamChoice type: object DeltaMessage: additionalProperties: false description: A message delta. properties: role: anyOf: - type: string - type: "null" default: null description: The role of the author of this message title: Role content: anyOf: - type: string - type: "null" default: null description: The contents of the chunk message title: Content internal_content: anyOf: - $ref: "#/components/schemas/InternalContent" - type: "null" default: null description: >- Internal model content (currently token_id) when include_internal_content=true. reasoning_content: anyOf: - type: string - type: "null" default: null description: >- The reasoning or thinking process generated by the model. This field is only available for certain reasoning models (GLM 4.5, GLM 4.5 Air, GPT OSS 120B, GPT OSS 20B) and contains the model's internal reasoning that would otherwise appear in `` tags within the content field. title: Reasoning Content tool_calls: anyOf: - items: $ref: "#/components/schemas/ChatCompletionMessageToolCall" type: array - type: "null" default: null title: Tool Calls title: DeltaMessage type: object InternalContent: additionalProperties: false description: >- Fireworks extension: internal model content surfaced inside the streaming delta when ``stream_options.include_internal_content=true``. properties: token_id: anyOf: - items: type: integer type: array - type: "null" default: null description: >- Token IDs for this chunk's delta (when include_internal_content=true) title: Token Id title: InternalContent type: object AnthropicMessage: examples: - content: - citations: null text: Hi! How can I help you today? type: text id: msg_013Zva2CMHLNnXjNJJKqJ2EF model: claude-opus-4-6 role: assistant stop_reason: end_turn stop_sequence: null type: message properties: id: description: |- Unique object identifier. The format and length of IDs may change over time. examples: - msg_013Zva2CMHLNnXjNJJKqJ2EF title: Id type: string type: const: message default: message description: |- Object type. For Messages, this is always `"message"`. title: Type type: string role: const: assistant default: assistant description: |- Conversational role of the generated message. This will always be `"assistant"`. title: Role type: string content: description: >- Content generated by the model. This is an array of content blocks, each of which has a `type` that determines its shape. Example: ```json [{"type": "text", "text": "Hi, I'm here to help."}] ``` If the request input `messages` ended with an `assistant` turn, then the response `content` will continue directly from that last turn. You can use this to constrain the model's output. For example, if the input `messages` were: ```json [ {"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"}, {"role": "assistant", "content": "The best answer is ("} ] ``` Then the response `content` might be: ```json [{"type": "text", "text": "B)"}] ``` examples: - - citations: null text: Hi! How can I help you today? type: text items: $ref: "#/components/schemas/AnthropicContentBlock" title: Content type: array model: $ref: "#/components/schemas/AnthropicModel" stop_reason: anyOf: - $ref: "#/components/schemas/AnthropicStopReason" - type: "null" description: >- The reason that the model stopped. This may be one the following values: * `"end_turn"`: the model reached a natural stopping point * `"max_tokens"`: the model exceeded the requested `max_tokens` or the model's maximum * `"stop_sequence"`: one of your provided custom `stop_sequences` was generated * `"tool_use"`: the model invoked one or more tools * `"pause_turn"`: the model paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue. * `"refusal"`: when streaming classifiers intervene to handle potential policy violations In non-streaming mode this value is always non-null. In streaming mode, it is null in the `message_start` event and non-null otherwise. title: Stop Reason stop_sequence: anyOf: - type: string - type: "null" default: null description: >- Which custom stop sequence was generated, if any. This value will be a non-null string if one of your custom stop sequences was generated. title: Stop Sequence raw_output: anyOf: - $ref: "#/components/schemas/AnthropicRawOutput" - type: "null" default: null required: - id - type - role - content - model - stop_reason - stop_sequence title: Message type: object x-stainless-python-custom-imports: - from .content_block import ContentBlock as ContentBlock AnthropicContentBlock: discriminator: mapping: redacted_thinking: "#/components/schemas/AnthropicResponseRedactedThinkingBlock" text: "#/components/schemas/AnthropicResponseTextBlock" thinking: "#/components/schemas/AnthropicResponseThinkingBlock" tool_use: "#/components/schemas/AnthropicResponseToolUseBlock" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicResponseTextBlock" - $ref: "#/components/schemas/AnthropicResponseThinkingBlock" - $ref: "#/components/schemas/AnthropicResponseRedactedThinkingBlock" - $ref: "#/components/schemas/AnthropicResponseToolUseBlock" AnthropicResponseTextBlock: properties: citations: anyOf: - items: discriminator: mapping: char_location: "#/components/schemas/AnthropicResponseCharLocationCitation" content_block_location: >- #/components/schemas/AnthropicResponseContentBlockLocationCitation page_location: "#/components/schemas/AnthropicResponsePageLocationCitation" search_result_location: >- #/components/schemas/AnthropicResponseSearchResultLocationCitation web_search_result_location: >- #/components/schemas/AnthropicResponseWebSearchResultLocationCitation propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicResponseCharLocationCitation" - $ref: "#/components/schemas/AnthropicResponsePageLocationCitation" - $ref: >- #/components/schemas/AnthropicResponseContentBlockLocationCitation - $ref: >- #/components/schemas/AnthropicResponseWebSearchResultLocationCitation - $ref: >- #/components/schemas/AnthropicResponseSearchResultLocationCitation type: array - type: "null" default: null description: >- Citations supporting the text block. The type of citation returned will depend on the type of document being cited. Citing a PDF results in `page_location`, plain text results in `char_location`, and content document results in `content_block_location`. title: Citations text: maxLength: 5000000 minLength: 0 title: Text type: string type: const: text default: text title: Type type: string required: - citations - text - type title: ResponseTextBlock type: object AnthropicResponseCharLocationCitation: properties: cited_text: title: Cited Text type: string document_index: minimum: 0 title: Document Index type: integer document_title: anyOf: - type: string - type: "null" title: Document Title end_char_index: title: End Char Index type: integer file_id: anyOf: - type: string - type: "null" default: null title: File Id start_char_index: minimum: 0 title: Start Char Index type: integer type: const: char_location default: char_location title: Type type: string required: - cited_text - document_index - document_title - end_char_index - file_id - start_char_index - type title: ResponseCharLocationCitation type: object AnthropicResponsePageLocationCitation: properties: cited_text: title: Cited Text type: string document_index: minimum: 0 title: Document Index type: integer document_title: anyOf: - type: string - type: "null" title: Document Title end_page_number: title: End Page Number type: integer file_id: anyOf: - type: string - type: "null" default: null title: File Id start_page_number: minimum: 1 title: Start Page Number type: integer type: const: page_location default: page_location title: Type type: string required: - cited_text - document_index - document_title - end_page_number - file_id - start_page_number - type title: ResponsePageLocationCitation type: object AnthropicResponseContentBlockLocationCitation: properties: cited_text: title: Cited Text type: string document_index: minimum: 0 title: Document Index type: integer document_title: anyOf: - type: string - type: "null" title: Document Title end_block_index: title: End Block Index type: integer file_id: anyOf: - type: string - type: "null" default: null title: File Id start_block_index: minimum: 0 title: Start Block Index type: integer type: const: content_block_location default: content_block_location title: Type type: string required: - cited_text - document_index - document_title - end_block_index - file_id - start_block_index - type title: ResponseContentBlockLocationCitation type: object AnthropicResponseWebSearchResultLocationCitation: properties: cited_text: title: Cited Text type: string encrypted_index: title: Encrypted Index type: string title: anyOf: - maxLength: 512 type: string - type: "null" title: Title type: const: web_search_result_location default: web_search_result_location title: Type type: string url: title: Url type: string required: - cited_text - encrypted_index - title - type - url title: ResponseWebSearchResultLocationCitation type: object AnthropicResponseSearchResultLocationCitation: properties: cited_text: title: Cited Text type: string end_block_index: title: End Block Index type: integer search_result_index: minimum: 0 title: Search Result Index type: integer source: title: Source type: string start_block_index: minimum: 0 title: Start Block Index type: integer title: anyOf: - type: string - type: "null" title: Title type: const: search_result_location default: search_result_location title: Type type: string required: - cited_text - end_block_index - search_result_index - source - start_block_index - title - type title: ResponseSearchResultLocationCitation type: object AnthropicResponseThinkingBlock: properties: signature: title: Signature type: string thinking: title: Thinking type: string type: const: thinking default: thinking title: Type type: string required: - signature - thinking - type title: ResponseThinkingBlock type: object AnthropicResponseRedactedThinkingBlock: properties: data: title: Data type: string type: const: redacted_thinking default: redacted_thinking title: Type type: string required: - data - type title: ResponseRedactedThinkingBlock type: object AnthropicResponseToolUseBlock: properties: id: pattern: ^[a-zA-Z0-9_-]+$ title: Id type: string input: additionalProperties: true title: Input type: object name: minLength: 1 title: Name type: string type: const: tool_use default: tool_use title: Type type: string required: - id - input - name - type title: ResponseToolUseBlock type: object AnthropicBase64PDFSource: additionalProperties: false properties: data: format: byte title: Data type: string media_type: const: application/pdf title: Media Type type: string type: const: base64 title: Type type: string required: - data - media_type - type title: Base64PDFSource type: object AnthropicPlainTextSource: additionalProperties: false properties: data: title: Data type: string media_type: const: text/plain title: Media Type type: string type: const: text title: Type type: string required: - data - media_type - type title: PlainTextSource type: object AnthropicModel: type: string title: Model description: >- The model that will complete your prompt. See the [Fireworks Model Library](https://app.fireworks.ai/models) for available models. AnthropicStopReason: enum: - end_turn - max_tokens - stop_sequence - tool_use - pause_turn - refusal type: string AnthropicErrorResponse: properties: error: discriminator: mapping: api_error: "#/components/schemas/AnthropicAPIError" authentication_error: "#/components/schemas/AnthropicAuthenticationError" billing_error: "#/components/schemas/AnthropicBillingError" invalid_request_error: "#/components/schemas/AnthropicInvalidRequestError" not_found_error: "#/components/schemas/AnthropicNotFoundError" overloaded_error: "#/components/schemas/AnthropicOverloadedError" permission_error: "#/components/schemas/AnthropicPermissionError" rate_limit_error: "#/components/schemas/AnthropicRateLimitError" timeout_error: "#/components/schemas/AnthropicGatewayTimeoutError" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicInvalidRequestError" - $ref: "#/components/schemas/AnthropicAuthenticationError" - $ref: "#/components/schemas/AnthropicBillingError" - $ref: "#/components/schemas/AnthropicPermissionError" - $ref: "#/components/schemas/AnthropicNotFoundError" - $ref: "#/components/schemas/AnthropicRateLimitError" - $ref: "#/components/schemas/AnthropicGatewayTimeoutError" - $ref: "#/components/schemas/AnthropicAPIError" - $ref: "#/components/schemas/AnthropicOverloadedError" title: Error request_id: anyOf: - type: string - type: "null" default: null title: Request Id type: const: error default: error title: Type type: string required: - error - request_id - type title: ErrorResponse type: object AnthropicInvalidRequestError: properties: message: default: Invalid request title: Message type: string type: const: invalid_request_error default: invalid_request_error title: Type type: string required: - message - type title: InvalidRequestError type: object AnthropicAuthenticationError: properties: message: default: Authentication error title: Message type: string type: const: authentication_error default: authentication_error title: Type type: string required: - message - type title: AuthenticationError type: object AnthropicBillingError: properties: message: default: Billing error title: Message type: string type: const: billing_error default: billing_error title: Type type: string required: - message - type title: BillingError type: object AnthropicPermissionError: properties: message: default: Permission denied title: Message type: string type: const: permission_error default: permission_error title: Type type: string required: - message - type title: PermissionError type: object AnthropicNotFoundError: properties: message: default: Not found title: Message type: string type: const: not_found_error default: not_found_error title: Type type: string required: - message - type title: NotFoundError type: object AnthropicRateLimitError: properties: message: default: Rate limited title: Message type: string type: const: rate_limit_error default: rate_limit_error title: Type type: string required: - message - type title: RateLimitError type: object AnthropicGatewayTimeoutError: properties: message: default: Request timeout title: Message type: string type: const: timeout_error default: timeout_error title: Type type: string required: - message - type title: GatewayTimeoutError type: object AnthropicAPIError: properties: message: default: Internal server error title: Message type: string type: const: api_error default: api_error title: Type type: string required: - message - type title: APIError type: object AnthropicOverloadedError: properties: message: default: Overloaded title: Message type: string type: const: overloaded_error default: overloaded_error title: Type type: string required: - message - type title: OverloadedError type: object AnthropicCreateMessageParams: additionalProperties: false example: max_tokens: 1024 messages: - content: Hello, world role: user model: claude-opus-4-6 properties: model: $ref: "#/components/schemas/AnthropicModel" messages: description: >- Input messages. Models are trained to operate on alternating `user` and `assistant` conversational turns. When creating a new `Message`, you specify the prior conversational turns with the `messages` parameter, and the model then generates the next `Message` in the conversation. Consecutive `user` or `assistant` turns in your request will be combined into a single turn. Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages. If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. Example with a single `user` message: ```json [{"role": "user", "content": "Hello"}] ``` Example with multiple conversational turns: ```json [ {"role": "user", "content": "Hello there."}, {"role": "assistant", "content": "Hi, I'm here to help. How can I help you?"}, {"role": "user", "content": "Can you explain LLMs in plain English?"}, ] ``` Example with a partially-filled response from the model: ```json [ {"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"}, {"role": "assistant", "content": "The best answer is ("}, ] ``` Each input message `content` may be either a single `string` or an array of content blocks, where each block has a specific `type`. Using a `string` for `content` is shorthand for an array of one content block of type `"text"`. The following input messages are equivalent: ```json {"role": "user", "content": "Hello"} ``` ```json {"role": "user", "content": [{"type": "text", "text": "Hello"}]} ``` See [input examples](https://docs.claude.com/en/api/messages-examples). Note that if you want to include a [system prompt](/guides/querying-text-models), you can use the top-level `system` parameter — there is no `"system"` role for input messages in the Messages API. There is a limit of 100,000 messages in a single request. items: $ref: "#/components/schemas/AnthropicInputMessage" title: Messages type: array max_tokens: description: >- The maximum number of tokens to generate before stopping. Note that models may stop _before_ reaching this maximum. This parameter only specifies the absolute maximum number of tokens to generate. Different models have different maximum values for this parameter. See [models](https://app.fireworks.ai/models) for details. examples: - 1024 minimum: 1 title: Max Tokens type: integer metadata: $ref: "#/components/schemas/AnthropicMetadata" description: An object describing metadata about the request. output_config: $ref: "#/components/schemas/AnthropicOutputConfig" description: >- Configuration options for the model's output, such as the output format. stop_sequences: description: >- Custom text sequences that will cause the model to stop generating. Models will normally stop when they have naturally completed their turn, which will result in a response `stop_reason` of `"end_turn"`. If you want the model to stop generating when it encounters custom strings of text, you can use the `stop_sequences` parameter. If the model encounters one of the custom sequences, the response `stop_reason` value will be `"stop_sequence"` and the response `stop_sequence` value will contain the matched stop sequence. items: type: string title: Stop Sequences type: array stream: description: >- Whether to incrementally stream the response using server-sent events. See [streaming](/guides/querying-text-models) for details. title: Stream type: boolean system: anyOf: - type: string x-stainless-skip: - go - cli - items: $ref: "#/components/schemas/AnthropicRequestTextBlock" type: array description: >- System prompt. A system prompt is a way of providing context and instructions to the model, such as specifying a particular goal or role. See the [guide to system prompts](/guides/querying-text-models). examples: - - text: Today's date is 2024-06-01. type: text - Today's date is 2023-01-01. title: System temperature: description: >- Amount of randomness injected into the response. Defaults to `1.0`. Ranges from `0.0` to `1.0`. Use `temperature` closer to `0.0` for analytical / multiple choice, and closer to `1.0` for creative and generative tasks. Note that even with `temperature` of `0.0`, the results will not be fully deterministic. examples: - 1 maximum: 1 minimum: 0 title: Temperature type: number thinking: $ref: "#/components/schemas/AnthropicThinkingConfigParam" tool_choice: $ref: "#/components/schemas/AnthropicToolChoice" tools: description: >- Definitions of tools that the model may use. If you include `tools` in your API request, the model may return `tool_use` content blocks that represent the model's use of those tools. You can then run those tools using the tool input generated by the model and then optionally return results back to the model using `tool_result` content blocks. Each tool definition includes: * `name`: Name of the tool. * `description`: Optional, but strongly-recommended description of the tool. * `input_schema`: [JSON schema](https://json-schema.org/draft/2020-12) for the tool `input` shape that the model will produce in `tool_use` output content blocks. For example, if you defined `tools` as: ```json [ { "name": "get_stock_price", "description": "Get the current stock price for a given ticker symbol.", "input_schema": { "type": "object", "properties": { "ticker": { "type": "string", "description": "The stock ticker symbol, e.g. AAPL for Apple Inc." } }, "required": ["ticker"] } } ] ``` And then asked the model "What's the S&P 500 at today?", the model might produce `tool_use` content blocks in the response like this: ```json [ { "type": "tool_use", "id": "toolu_01D7FLrfh4GYq7yT1ULFeyMV", "name": "get_stock_price", "input": { "ticker": "^GSPC" } } ] ``` You might then run your `get_stock_price` tool with `{"ticker": "^GSPC"}` as an input, and return the following back to the model in a subsequent `user` message: ```json [ { "type": "tool_result", "tool_use_id": "toolu_01D7FLrfh4GYq7yT1ULFeyMV", "content": "259.75 USD" } ] ``` Tools can be used for workflows that include running client-side tools and functions, or more generally whenever you want the model to produce a particular JSON structure of output. See the [guide](/guides/function-calling) for more details. examples: - description: Get the current weather in a given location input_schema: properties: location: description: The city and state, e.g. San Francisco, CA type: string unit: description: Unit for the output - one of (celsius, fahrenheit) type: string required: - location type: object name: get_weather items: oneOf: - $ref: "#/components/schemas/AnthropicTool" title: Tools type: array top_k: description: >- Only sample from the top K options for each subsequent token. Used to remove "long tail" low probability responses. [Learn more technical details here](https://towardsdatascience.com/how-to-sample-from-language-models-682bceb97277). Recommended for advanced use cases only. You usually only need to use `temperature`. examples: - 5 minimum: 0 title: Top K type: integer top_p: description: >- Use nucleus sampling. In nucleus sampling, we compute the cumulative distribution over all the options for each subsequent token in decreasing probability order and cut it off once it reaches a particular probability specified by `top_p`. You should either alter `temperature` or `top_p`, but not both. Recommended for advanced use cases only. You usually only need to use `temperature`. examples: - 0.7 maximum: 1 minimum: 0 title: Top P type: number raw_output: anyOf: - type: boolean - type: "null" title: Raw Output description: Return raw output from the model. default: false required: - model - messages title: CreateMessageParams type: object AnthropicInputMessage: additionalProperties: false properties: content: anyOf: - type: string x-stainless-skip: - go - cli - items: $ref: "#/components/schemas/AnthropicInputContentBlock" type: array example: - type: text text: What is a quaternion? title: Content role: enum: - user - assistant title: Role type: string required: - content - role title: InputMessage type: object discriminator: propertyName: role AnthropicInputContentBlock: discriminator: mapping: document: "#/components/schemas/AnthropicRequestDocumentBlock" image: "#/components/schemas/AnthropicRequestImageBlock" redacted_thinking: "#/components/schemas/AnthropicRequestRedactedThinkingBlock" text: "#/components/schemas/AnthropicRequestTextBlock" thinking: "#/components/schemas/AnthropicRequestThinkingBlock" tool_result: "#/components/schemas/AnthropicRequestToolResultBlock" tool_use: "#/components/schemas/AnthropicRequestToolUseBlock" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicRequestTextBlock" description: Regular text content. - $ref: "#/components/schemas/AnthropicRequestImageBlock" description: >- Image content specified directly as base64 data or as a reference via a URL. - $ref: "#/components/schemas/AnthropicRequestDocumentBlock" description: >- Document content, either specified directly as base64 data, as text, or as a reference via a URL. - $ref: "#/components/schemas/AnthropicRequestThinkingBlock" description: A block specifying internal thinking by the model. - $ref: "#/components/schemas/AnthropicRequestRedactedThinkingBlock" description: A block specifying internal, redacted thinking by the model. - $ref: "#/components/schemas/AnthropicRequestToolUseBlock" description: A block indicating a tool use by the model. - $ref: "#/components/schemas/AnthropicRequestToolResultBlock" description: A block specifying the results of a tool use by the model. x-stainless-python-extend-union: - ContentBlock x-stainless-python-extend-union-imports: - from .content_block import ContentBlock x-stainless-go-variant-constructor: naming: new_{variant}_block AnthropicRequestTextBlock: additionalProperties: false properties: cache_control: anyOf: - discriminator: mapping: ephemeral: "#/components/schemas/AnthropicCacheControlEphemeral" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicCacheControlEphemeral" - type: "null" description: Create a cache control breakpoint at this content block. title: Cache Control citations: anyOf: - items: discriminator: mapping: char_location: "#/components/schemas/AnthropicRequestCharLocationCitation" content_block_location: >- #/components/schemas/AnthropicRequestContentBlockLocationCitation page_location: "#/components/schemas/AnthropicRequestPageLocationCitation" search_result_location: >- #/components/schemas/AnthropicRequestSearchResultLocationCitation web_search_result_location: >- #/components/schemas/AnthropicRequestWebSearchResultLocationCitation propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicRequestCharLocationCitation" - $ref: "#/components/schemas/AnthropicRequestPageLocationCitation" - $ref: >- #/components/schemas/AnthropicRequestContentBlockLocationCitation - $ref: >- #/components/schemas/AnthropicRequestWebSearchResultLocationCitation - $ref: >- #/components/schemas/AnthropicRequestSearchResultLocationCitation type: array - type: "null" title: Citations text: minLength: 1 title: Text type: string type: const: text title: Type type: string required: - text - type title: RequestTextBlock type: object AnthropicCacheControlEphemeral: additionalProperties: false properties: ttl: description: |- The time-to-live for the cache control breakpoint. This may be one the following values: - `5m`: 5 minutes - `1h`: 1 hour Defaults to `5m`. enum: - 5m - 1h title: Ttl type: string x-stainless-renameMap: ttl_5m: 5m ttl_1h: 1h type: const: ephemeral title: Type type: string required: - type title: CacheControlEphemeral type: object x-stainless-go-constant-constructor: true AnthropicRequestCharLocationCitation: additionalProperties: false properties: cited_text: title: Cited Text type: string document_index: minimum: 0 title: Document Index type: integer document_title: anyOf: - maxLength: 255 minLength: 1 type: string - type: "null" title: Document Title end_char_index: title: End Char Index type: integer start_char_index: minimum: 0 title: Start Char Index type: integer type: const: char_location title: Type type: string required: - cited_text - document_index - document_title - end_char_index - start_char_index - type title: RequestCharLocationCitation type: object AnthropicRequestPageLocationCitation: additionalProperties: false properties: cited_text: title: Cited Text type: string document_index: minimum: 0 title: Document Index type: integer document_title: anyOf: - maxLength: 255 minLength: 1 type: string - type: "null" title: Document Title end_page_number: title: End Page Number type: integer start_page_number: minimum: 1 title: Start Page Number type: integer type: const: page_location title: Type type: string required: - cited_text - document_index - document_title - end_page_number - start_page_number - type title: RequestPageLocationCitation type: object AnthropicRequestContentBlockLocationCitation: additionalProperties: false properties: cited_text: title: Cited Text type: string document_index: minimum: 0 title: Document Index type: integer document_title: anyOf: - maxLength: 255 minLength: 1 type: string - type: "null" title: Document Title end_block_index: title: End Block Index type: integer start_block_index: minimum: 0 title: Start Block Index type: integer type: const: content_block_location title: Type type: string required: - cited_text - document_index - document_title - end_block_index - start_block_index - type title: RequestContentBlockLocationCitation type: object AnthropicRequestWebSearchResultLocationCitation: additionalProperties: false properties: cited_text: title: Cited Text type: string encrypted_index: title: Encrypted Index type: string title: anyOf: - maxLength: 512 minLength: 1 type: string - type: "null" title: Title type: const: web_search_result_location title: Type type: string url: maxLength: 2048 minLength: 1 title: Url type: string required: - cited_text - encrypted_index - title - type - url title: RequestWebSearchResultLocationCitation type: object AnthropicRequestSearchResultLocationCitation: additionalProperties: false properties: cited_text: title: Cited Text type: string end_block_index: title: End Block Index type: integer search_result_index: minimum: 0 title: Search Result Index type: integer source: title: Source type: string start_block_index: minimum: 0 title: Start Block Index type: integer title: anyOf: - type: string - type: "null" title: Title type: const: search_result_location title: Type type: string required: - cited_text - end_block_index - search_result_index - source - start_block_index - title - type title: RequestSearchResultLocationCitation type: object AnthropicRequestImageBlock: additionalProperties: false properties: cache_control: anyOf: - discriminator: mapping: ephemeral: "#/components/schemas/AnthropicCacheControlEphemeral" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicCacheControlEphemeral" - type: "null" description: Create a cache control breakpoint at this content block. title: Cache Control source: discriminator: mapping: base64: "#/components/schemas/AnthropicBase64ImageSource" url: "#/components/schemas/AnthropicURLImageSource" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicBase64ImageSource" - $ref: "#/components/schemas/AnthropicURLImageSource" title: Source type: const: image title: Type type: string required: - source - type title: RequestImageBlock type: object AnthropicBase64ImageSource: additionalProperties: false properties: data: format: byte title: Data type: string media_type: enum: - image/jpeg - image/png - image/gif - image/webp title: Media Type type: string type: const: base64 title: Type type: string required: - data - media_type - type title: Base64ImageSource type: object AnthropicURLImageSource: additionalProperties: false properties: type: const: url title: Type type: string url: title: Url type: string required: - type - url title: URLImageSource type: object AnthropicRequestDocumentBlock: additionalProperties: false properties: cache_control: anyOf: - discriminator: mapping: ephemeral: "#/components/schemas/AnthropicCacheControlEphemeral" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicCacheControlEphemeral" - type: "null" description: Create a cache control breakpoint at this content block. title: Cache Control citations: anyOf: - $ref: "#/components/schemas/AnthropicRequestCitationsConfig" - type: "null" context: anyOf: - minLength: 1 type: string - type: "null" title: Context source: discriminator: mapping: base64: "#/components/schemas/AnthropicBase64PDFSource" content: "#/components/schemas/AnthropicContentBlockSource" text: "#/components/schemas/AnthropicPlainTextSource" url: "#/components/schemas/AnthropicURLPDFSource" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicBase64PDFSource" - $ref: "#/components/schemas/AnthropicPlainTextSource" - $ref: "#/components/schemas/AnthropicContentBlockSource" - $ref: "#/components/schemas/AnthropicURLPDFSource" title: Source title: anyOf: - maxLength: 500 minLength: 1 type: string - type: "null" title: Title type: const: document title: Type type: string required: - source - type title: RequestDocumentBlock type: object AnthropicRequestCitationsConfig: additionalProperties: false properties: enabled: title: Enabled type: boolean title: RequestCitationsConfig type: object AnthropicContentBlockSource: additionalProperties: false properties: content: anyOf: - type: string - items: discriminator: mapping: image: "#/components/schemas/AnthropicRequestImageBlock" text: "#/components/schemas/AnthropicRequestTextBlock" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicRequestTextBlock" - $ref: "#/components/schemas/AnthropicRequestImageBlock" x-stainless-naming: go: type_name: ContentBlockSourceContentItem title: content_block_source_content_item type: array title: content_block_source_content title: Content type: const: content title: Type type: string required: - content - type title: ContentBlockSource type: object AnthropicURLPDFSource: additionalProperties: false properties: type: const: url title: Type type: string url: title: Url type: string required: - type - url title: URLPDFSource type: object AnthropicRequestThinkingBlock: additionalProperties: false properties: signature: title: Signature type: string thinking: title: Thinking type: string type: const: thinking title: Type type: string required: - signature - thinking - type title: RequestThinkingBlock type: object AnthropicRequestRedactedThinkingBlock: additionalProperties: false properties: data: title: Data type: string type: const: redacted_thinking title: Type type: string required: - data - type title: RequestRedactedThinkingBlock type: object AnthropicRequestToolUseBlock: additionalProperties: false properties: cache_control: anyOf: - discriminator: mapping: ephemeral: "#/components/schemas/AnthropicCacheControlEphemeral" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicCacheControlEphemeral" - type: "null" description: Create a cache control breakpoint at this content block. title: Cache Control id: pattern: ^[a-zA-Z0-9_-]+$ title: Id type: string input: additionalProperties: true title: Input type: object name: maxLength: 200 minLength: 1 title: Name type: string type: const: tool_use title: Type type: string required: - id - input - name - type title: RequestToolUseBlock type: object AnthropicRequestToolResultBlock: additionalProperties: false properties: cache_control: anyOf: - discriminator: mapping: ephemeral: "#/components/schemas/AnthropicCacheControlEphemeral" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicCacheControlEphemeral" - type: "null" description: Create a cache control breakpoint at this content block. title: Cache Control content: anyOf: - type: string x-stainless-skip: - go - cli - items: discriminator: mapping: document: "#/components/schemas/AnthropicRequestDocumentBlock" image: "#/components/schemas/AnthropicRequestImageBlock" text: "#/components/schemas/AnthropicRequestTextBlock" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicRequestTextBlock" - $ref: "#/components/schemas/AnthropicRequestImageBlock" - $ref: "#/components/schemas/AnthropicRequestDocumentBlock" title: Block type: array x-stainless-naming: python: type_name: Content ruby: type_name: Content php: type_name: Content title: Content is_error: title: Is Error type: boolean tool_use_id: pattern: ^[a-zA-Z0-9_-]+$ title: Tool Use Id type: string type: const: tool_result title: Type type: string required: - tool_use_id - type title: RequestToolResultBlock type: object AnthropicMetadata: additionalProperties: false properties: user_id: anyOf: - maxLength: 256 type: string - type: "null" description: >- An external identifier for the user who is associated with the request. This should be a uuid, hash value, or other opaque identifier. This id may be used to help detect abuse. Do not include any identifying information such as name, email address, or phone number. examples: - 13803d75-b4b5-4c3e-b2a2-6f21399b021b title: User Id title: Metadata type: object AnthropicOutputConfig: additionalProperties: false properties: effort: anyOf: - $ref: "#/components/schemas/AnthropicEffortLevel" - type: "null" description: >- How much effort the model should put into its response. Higher effort levels may result in more thorough analysis but take longer. Valid values are `low`, `medium`, `high`, or `max`. **Fireworks behavior:** The `effort` level is converted to Fireworks [`reasoning_effort`](/api-reference/post-chatcompletions). Values `low`, `medium`, and `high` map directly. `max` is mapped to `high` since Fireworks does not have a `max` effort level. format: anyOf: - $ref: "#/components/schemas/AnthropicJsonOutputFormat" - type: "null" description: >- A schema to specify the model's output format in responses. See [structured outputs](/structured-responses/structured-output-grammar-based) title: OutputConfig type: object AnthropicEffortLevel: description: All possible effort levels. enum: - low - medium - high - max title: EffortLevel type: string AnthropicJsonOutputFormat: additionalProperties: false properties: schema: additionalProperties: true description: The JSON schema of the format title: Schema type: object type: const: json_schema title: Type type: string required: - schema - type title: JsonOutputFormat type: object AnthropicThinkingConfigParam: description: >- Configuration for enabling the model's extended thinking. When enabled, responses include `thinking` content blocks showing the model's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit. See [reasoning](/guides/reasoning) for details. **Note:** The `adaptive` thinking type is not supported yet. discriminator: mapping: adaptive: "#/components/schemas/AnthropicThinkingConfigAdaptive" disabled: "#/components/schemas/AnthropicThinkingConfigDisabled" enabled: "#/components/schemas/AnthropicThinkingConfigEnabled" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicThinkingConfigEnabled" - $ref: "#/components/schemas/AnthropicThinkingConfigDisabled" - $ref: "#/components/schemas/AnthropicThinkingConfigAdaptive" title: Thinking AnthropicThinkingConfigEnabled: additionalProperties: false properties: budget_tokens: description: >- Determines how many tokens the model can use for its internal reasoning process. Larger budgets can enable more thorough analysis for complex problems, improving response quality. Must be ≥1024 and less than `max_tokens`. See [reasoning](/guides/reasoning) for details. minimum: 1024 title: Budget Tokens type: integer type: const: enabled title: Type type: string required: - budget_tokens - type title: ThinkingConfigEnabled type: object AnthropicThinkingConfigDisabled: additionalProperties: false properties: type: const: disabled title: Type type: string required: - type title: ThinkingConfigDisabled type: object x-stainless-go-constant-constructor: true AnthropicThinkingConfigAdaptive: additionalProperties: false properties: type: const: adaptive title: Type type: string required: - type title: ThinkingConfigAdaptive type: object description: "**Not supported yet.**" AnthropicToolChoice: description: >- How the model should use the provided tools. The model can use a specific tool, any available tool, decide by itself, or not use tools at all. discriminator: mapping: any: "#/components/schemas/AnthropicToolChoiceAny" auto: "#/components/schemas/AnthropicToolChoiceAuto" none: "#/components/schemas/AnthropicToolChoiceNone" tool: "#/components/schemas/AnthropicToolChoiceTool" propertyName: type oneOf: - $ref: "#/components/schemas/AnthropicToolChoiceAuto" - $ref: "#/components/schemas/AnthropicToolChoiceAny" - $ref: "#/components/schemas/AnthropicToolChoiceTool" - $ref: "#/components/schemas/AnthropicToolChoiceNone" title: Tool Choice AnthropicToolChoiceAuto: additionalProperties: false description: The model will automatically decide whether to use tools. properties: disable_parallel_tool_use: description: >- Whether to disable parallel tool use. Defaults to `false`. If set to `true`, the model will output at most one tool use. title: Disable Parallel Tool Use type: boolean type: const: auto title: Type type: string required: - type title: ToolChoiceAuto type: object AnthropicToolChoiceAny: additionalProperties: false description: The model will use any available tools. properties: disable_parallel_tool_use: description: >- Whether to disable parallel tool use. Defaults to `false`. If set to `true`, the model will output exactly one tool use. title: Disable Parallel Tool Use type: boolean type: const: any title: Type type: string required: - type title: ToolChoiceAny type: object AnthropicToolChoiceTool: additionalProperties: false description: The model will use the specified tool with `tool_choice.name`. properties: disable_parallel_tool_use: description: >- Whether to disable parallel tool use. Defaults to `false`. If set to `true`, the model will output exactly one tool use. title: Disable Parallel Tool Use type: boolean name: description: The name of the tool to use. title: Name type: string type: const: tool title: Type type: string required: - name - type title: ToolChoiceTool type: object AnthropicToolChoiceNone: additionalProperties: false description: The model will not be allowed to use tools. properties: type: const: none title: Type type: string required: - type title: ToolChoiceNone type: object x-stainless-go-constant-constructor: true AnthropicTool: additionalProperties: false properties: type: anyOf: - type: "null" - const: custom type: string title: Type description: description: >- Description of what this tool does. Tool descriptions should be as detailed as possible. The more information that the model has about what the tool is and how to use it, the better it will perform. You can use natural language descriptions to reinforce important aspects of the tool input JSON schema. examples: - Get the current weather in a given location title: Description type: string name: description: >- Name of the tool. This is how the tool will be called by the model and in `tool_use` blocks. maxLength: 128 minLength: 1 pattern: ^[a-zA-Z0-9_-]{1,128}$ title: Name type: string input_schema: $ref: "#/components/schemas/AnthropicInputSchema" description: >- [JSON schema](https://json-schema.org/draft/2020-12) for this tool's input. This defines the shape of the `input` that your tool accepts and that the model will produce. examples: - properties: location: description: The city and state, e.g. San Francisco, CA type: string unit: description: Unit for the output - one of (celsius, fahrenheit) type: string required: - location type: object x-stainless-skip: - cli strict: description: When true, guarantees schema validation on tool names and inputs title: Strict type: boolean required: - name - input_schema title: Tool type: object AnthropicInputSchema: additionalProperties: true properties: properties: anyOf: - additionalProperties: true type: object - type: "null" title: Properties required: anyOf: - items: type: string type: array - type: "null" title: Required type: const: object title: Type type: string required: - type title: InputSchema type: object AnthropicRawOutput: additionalProperties: false description: >- Fireworks extension that returns low-level details of what the model sees, including the formatted prompt and function calls. properties: prompt_fragments: description: >- Pieces of the prompt (like individual messages) before truncation and concatenation. Depending on prompt_truncate_len some of the messages might be dropped. Contains a mix of strings to be tokenized and individual tokens (if dictated by the conversation template) items: anyOf: - type: string - type: integer title: Prompt Fragments type: array prompt_token_ids: description: Fully processed prompt as seen by the model items: type: integer title: Prompt Token Ids type: array completion: description: >- Raw completion produced by the model before any tool calls are parsed title: Completion type: string completion_token_ids: anyOf: - items: type: integer type: array - type: "null" default: null description: Token IDs for the raw completion title: Completion Token Ids images: anyOf: - items: type: string type: array - type: "null" default: null description: Images in the prompt title: Images grammar: anyOf: - type: string - type: "null" default: null description: >- Grammar used for constrained decoding, can be either user provided (directly or JSON schema) or inferred by the chat template title: Grammar required: - prompt_fragments - prompt_token_ids - completion title: RawOutput type: object AnthropicMessageStartEvent: type: object properties: type: const: message_start type: string message: $ref: "#/components/schemas/AnthropicMessage" required: - type - message AnthropicContentBlockStartEvent: type: object properties: type: const: content_block_start type: string index: type: integer content_block: $ref: "#/components/schemas/AnthropicContentBlock" required: - type - index - content_block AnthropicContentBlockDeltaEvent: type: object properties: type: const: content_block_delta type: string index: type: integer delta: type: object properties: type: type: string enum: - text_delta - thinking_delta - signature_delta - input_json_delta - citations_delta text: type: string thinking: type: string signature: type: string partial_json: type: string required: - type required: - type - index - delta AnthropicContentBlockStopEvent: type: object properties: type: const: content_block_stop type: string index: type: integer required: - type - index AnthropicMessageDeltaEvent: type: object properties: type: const: message_delta type: string delta: type: object properties: stop_reason: $ref: "#/components/schemas/AnthropicStopReason" stop_sequence: anyOf: - type: string - type: "null" usage: type: object properties: output_tokens: type: integer required: - output_tokens required: - type - delta - usage AnthropicMessageStopEvent: type: object properties: type: const: message_stop type: string required: - type AnthropicPingEvent: type: object properties: type: const: ping type: string required: - type AnthropicMessageStreamEvent: discriminator: propertyName: type mapping: message_start: "#/components/schemas/AnthropicMessageStartEvent" content_block_start: "#/components/schemas/AnthropicContentBlockStartEvent" content_block_delta: "#/components/schemas/AnthropicContentBlockDeltaEvent" content_block_stop: "#/components/schemas/AnthropicContentBlockStopEvent" message_delta: "#/components/schemas/AnthropicMessageDeltaEvent" message_stop: "#/components/schemas/AnthropicMessageStopEvent" ping: "#/components/schemas/AnthropicPingEvent" oneOf: - $ref: "#/components/schemas/AnthropicMessageStartEvent" - $ref: "#/components/schemas/AnthropicContentBlockStartEvent" - $ref: "#/components/schemas/AnthropicContentBlockDeltaEvent" - $ref: "#/components/schemas/AnthropicContentBlockStopEvent" - $ref: "#/components/schemas/AnthropicMessageDeltaEvent" - $ref: "#/components/schemas/AnthropicMessageStopEvent" - $ref: "#/components/schemas/AnthropicPingEvent" description: A Server-Sent Event from the Messages API stream. x-tagGroups: - name: Gateway REST API tags: - gateway.openapi_AccountService - gateway.openapi_DeploymentService - gateway.openapi_Gateway - gateway.openapi_ModelService - gateway.openapi_TrainingService - gateway-extra.openapi_Gateway - name: Fireworks Responses API tags: - responses-harmonized.openapi_other - name: Fireworks Text Completion API tags: - text-completion.openapi_other - name: Fireworks AI Anthropic Compatible Messages API tags: - anthropic-messages.openapi_other