openapi: 3.1.0
info:
  title: Llama Stack Specification
  version: v1
  description: >-
    This is the specification of the Llama Stack that provides
                    a set of endpoints and their corresponding interfaces that are
    tailored to
                    best leverage Llama Models.
servers:
  - url: http://any-hosted-llama-stack.com
paths:
  /v1/datasetio/rows:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/PaginatedRowsResult'
      tags:
        - DatasetIO
      parameters:
        - name: dataset_id
          in: query
          required: true
          schema:
            type: string
        - name: rows_in_page
          in: query
          required: true
          schema:
            type: integer
        - name: page_token
          in: query
          required: false
          schema:
            type: string
        - name: filter_condition
          in: query
          required: false
          schema:
            type: string
    post:
      responses:
        '200':
          description: OK
      tags:
        - DatasetIO
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/AppendRowsRequest'
        required: true
  /v1/batch-inference/chat-completion:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/BatchChatCompletionResponse'
      tags:
        - BatchInference (Coming Soon)
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/BatchChatCompletionRequest'
        required: true
  /v1/batch-inference/completion:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/BatchCompletionResponse'
      tags:
        - BatchInference (Coming Soon)
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/BatchCompletionRequest'
        required: true
  /v1/post-training/job/cancel:
    post:
      responses:
        '200':
          description: OK
      tags:
        - PostTraining (Coming Soon)
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CancelTrainingJobRequest'
        required: true
  /v1/inference/chat-completion:
    post:
      responses:
        '200':
          description: >-
            If stream=False, returns a ChatCompletionResponse with the full completion.
            If stream=True, returns an SSE event stream of ChatCompletionResponseStreamChunk
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ChatCompletionResponse'
            text/event-stream:
              schema:
                $ref: '#/components/schemas/ChatCompletionResponseStreamChunk'
      tags:
        - Inference
      summary: >-
        Generate a chat completion for the given messages using the specified model.
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ChatCompletionRequest'
        required: true
  /v1/inference/completion:
    post:
      responses:
        '200':
          description: >-
            If stream=False, returns a CompletionResponse with the full completion.
            If stream=True, returns an SSE event stream of CompletionResponseStreamChunk
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CompletionResponse'
            text/event-stream:
              schema:
                $ref: '#/components/schemas/CompletionResponseStreamChunk'
      tags:
        - Inference
      summary: >-
        Generate a completion for the given content using the specified model.
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CompletionRequest'
        required: true
  /v1/agents:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AgentCreateResponse'
      tags:
        - Agents
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateAgentRequest'
        required: true
  /v1/agents/{agent_id}/session:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AgentSessionCreateResponse'
      tags:
        - Agents
      parameters:
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateAgentSessionRequest'
        required: true
  /v1/agents/{agent_id}/session/{session_id}/turn:
    post:
      responses:
        '200':
          description: >-
            A single turn in an interaction with an Agentic System. **OR** streamed
            agent turn completion response.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Turn'
            text/event-stream:
              schema:
                $ref: '#/components/schemas/AgentTurnResponseStreamChunk'
      tags:
        - Agents
      parameters:
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
        - name: session_id
          in: path
          required: true
          schema:
            type: string
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateAgentTurnRequest'
        required: true
  /v1/agents/{agent_id}:
    delete:
      responses:
        '200':
          description: OK
      tags:
        - Agents
      parameters:
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
  /v1/agents/{agent_id}/session/{session_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Session'
      tags:
        - Agents
      parameters:
        - name: session_id
          in: path
          required: true
          schema:
            type: string
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
        - name: turn_ids
          in: query
          required: false
          schema:
            type: array
            items:
              type: string
    delete:
      responses:
        '200':
          description: OK
      tags:
        - Agents
      parameters:
        - name: session_id
          in: path
          required: true
          schema:
            type: string
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
  /v1/inference/embeddings:
    post:
      responses:
        '200':
          description: >-
            An array of embeddings, one for each content. Each embedding is a list
            of floats. The dimensionality of the embedding is model-specific; you
            can check model metadata using /models/{model_id}
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/EmbeddingsResponse'
      tags:
        - Inference
      summary: >-
        Generate embeddings for content pieces using the specified model.
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/EmbeddingsRequest'
        required: true
  /v1/eval/tasks/{task_id}/evaluations:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/EvaluateResponse'
      tags:
        - Eval
      parameters:
        - name: task_id
          in: path
          required: true
          schema:
            type: string
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/EvaluateRowsRequest'
        required: true
  /v1/agents/{agent_id}/session/{session_id}/turn/{turn_id}/step/{step_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AgentStepResponse'
      tags:
        - Agents
      parameters:
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
        - name: session_id
          in: path
          required: true
          schema:
            type: string
        - name: turn_id
          in: path
          required: true
          schema:
            type: string
        - name: step_id
          in: path
          required: true
          schema:
            type: string
  /v1/agents/{agent_id}/session/{session_id}/turn/{turn_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Turn'
      tags:
        - Agents
      parameters:
        - name: agent_id
          in: path
          required: true
          schema:
            type: string
        - name: session_id
          in: path
          required: true
          schema:
            type: string
        - name: turn_id
          in: path
          required: true
          schema:
            type: string
  /v1/datasets/{dataset_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/Dataset'
                  - type: 'null'
      tags:
        - Datasets
      parameters:
        - name: dataset_id
          in: path
          required: true
          schema:
            type: string
    delete:
      responses:
        '200':
          description: OK
      tags:
        - Datasets
      parameters:
        - name: dataset_id
          in: path
          required: true
          schema:
            type: string
  /v1/eval-tasks/{eval_task_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/EvalTask'
                  - type: 'null'
      tags:
        - EvalTasks
      parameters:
        - name: eval_task_id
          in: path
          required: true
          schema:
            type: string
  /v1/models/{model_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/Model'
                  - type: 'null'
      tags:
        - Models
      parameters:
        - name: model_id
          in: path
          required: true
          schema:
            type: string
    delete:
      responses:
        '200':
          description: OK
      tags:
        - Models
      parameters:
        - name: model_id
          in: path
          required: true
          schema:
            type: string
  /v1/scoring-functions/{scoring_fn_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/ScoringFn'
                  - type: 'null'
      tags:
        - ScoringFunctions
      parameters:
        - name: scoring_fn_id
          in: path
          required: true
          schema:
            type: string
  /v1/shields/{identifier}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/Shield'
                  - type: 'null'
      tags:
        - Shields
      parameters:
        - name: identifier
          in: path
          required: true
          schema:
            type: string
  /v1/telemetry/traces/{trace_id}/spans/{span_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Span'
      tags:
        - Telemetry
      parameters:
        - name: trace_id
          in: path
          required: true
          schema:
            type: string
        - name: span_id
          in: path
          required: true
          schema:
            type: string
  /v1/telemetry/spans/{span_id}/tree:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QuerySpanTreeResponse'
      tags:
        - Telemetry
      parameters:
        - name: span_id
          in: path
          required: true
          schema:
            type: string
        - name: attributes_to_return
          in: query
          required: false
          schema:
            type: array
            items:
              type: string
        - name: max_depth
          in: query
          required: false
          schema:
            type: integer
  /v1/tools/{tool_name}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Tool'
      tags:
        - ToolGroups
      parameters:
        - name: tool_name
          in: path
          required: true
          schema:
            type: string
  /v1/toolgroups/{toolgroup_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ToolGroup'
      tags:
        - ToolGroups
      parameters:
        - name: toolgroup_id
          in: path
          required: true
          schema:
            type: string
    delete:
      responses:
        '200':
          description: OK
      tags:
        - ToolGroups
      summary: Unregister a tool group
      parameters:
        - name: toolgroup_id
          in: path
          required: true
          schema:
            type: string
  /v1/telemetry/traces/{trace_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Trace'
      tags:
        - Telemetry
      parameters:
        - name: trace_id
          in: path
          required: true
          schema:
            type: string
  /v1/post-training/job/artifacts:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/PostTrainingJobArtifactsResponse'
                  - type: 'null'
      tags:
        - PostTraining (Coming Soon)
      parameters:
        - name: job_uuid
          in: query
          required: true
          schema:
            type: string
  /v1/post-training/job/status:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/PostTrainingJobStatusResponse'
                  - type: 'null'
      tags:
        - PostTraining (Coming Soon)
      parameters:
        - name: job_uuid
          in: query
          required: true
          schema:
            type: string
  /v1/post-training/jobs:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListPostTrainingJobsResponse'
      tags:
        - PostTraining (Coming Soon)
      parameters: []
  /v1/vector-dbs/{vector_db_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/VectorDB'
                  - type: 'null'
      tags:
        - VectorDBs
      parameters:
        - name: vector_db_id
          in: path
          required: true
          schema:
            type: string
    delete:
      responses:
        '200':
          description: OK
      tags:
        - VectorDBs
      parameters:
        - name: vector_db_id
          in: path
          required: true
          schema:
            type: string
  /v1/health:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/HealthInfo'
      tags:
        - Inspect
      parameters: []
  /v1/tool-runtime/rag-tool/insert:
    post:
      responses:
        '200':
          description: OK
      tags:
        - ToolRuntime
      summary: >-
        Index documents so they can be used by the RAG system
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/InsertRequest'
        required: true
  /v1/vector-io/insert:
    post:
      responses:
        '200':
          description: OK
      tags:
        - VectorIO
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/InsertChunksRequest'
        required: true
  /v1/tool-runtime/invoke:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ToolInvocationResult'
      tags:
        - ToolRuntime
      summary: Run a tool with the given arguments
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/InvokeToolRequest'
        required: true
  /v1/eval/tasks/{task_id}/jobs/{job_id}:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                oneOf:
                  - $ref: '#/components/schemas/JobStatus'
                  - type: 'null'
      tags:
        - Eval
      parameters:
        - name: task_id
          in: path
          required: true
          schema:
            type: string
        - name: job_id
          in: path
          required: true
          schema:
            type: string
    delete:
      responses:
        '200':
          description: OK
      tags:
        - Eval
      parameters:
        - name: task_id
          in: path
          required: true
          schema:
            type: string
        - name: job_id
          in: path
          required: true
          schema:
            type: string
  /v1/eval/tasks/{task_id}/jobs/{job_id}/result:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/EvaluateResponse'
      tags:
        - Eval
      parameters:
        - name: job_id
          in: path
          required: true
          schema:
            type: string
        - name: task_id
          in: path
          required: true
          schema:
            type: string
  /v1/datasets:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListDatasetsResponse'
      tags:
        - Datasets
      parameters: []
    post:
      responses:
        '200':
          description: OK
      tags:
        - Datasets
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterDatasetRequest'
        required: true
  /v1/eval-tasks:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListEvalTasksResponse'
      tags:
        - EvalTasks
      parameters: []
    post:
      responses:
        '200':
          description: OK
      tags:
        - EvalTasks
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterEvalTaskRequest'
        required: true
  /v1/models:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListModelsResponse'
      tags:
        - Models
      parameters: []
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Model'
      tags:
        - Models
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterModelRequest'
        required: true
  /v1/inspect/providers:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListProvidersResponse'
      tags:
        - Inspect
      parameters: []
  /v1/inspect/routes:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListRoutesResponse'
      tags:
        - Inspect
      parameters: []
  /v1/tool-runtime/list-tools:
    get:
      responses:
        '200':
          description: OK
          content:
            application/jsonl:
              schema:
                $ref: '#/components/schemas/ToolDef'
      tags:
        - ToolRuntime
      parameters:
        - name: tool_group_id
          in: query
          required: false
          schema:
            type: string
        - name: mcp_endpoint
          in: query
          required: false
          schema:
            $ref: '#/components/schemas/URL'
  /v1/scoring-functions:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListScoringFunctionsResponse'
      tags:
        - ScoringFunctions
      parameters: []
    post:
      responses:
        '200':
          description: OK
      tags:
        - ScoringFunctions
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterScoringFunctionRequest'
        required: true
  /v1/shields:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListShieldsResponse'
      tags:
        - Shields
      parameters: []
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Shield'
      tags:
        - Shields
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterShieldRequest'
        required: true
  /v1/toolgroups:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListToolGroupsResponse'
      tags:
        - ToolGroups
      summary: List tool groups with optional provider
      parameters: []
    post:
      responses:
        '200':
          description: OK
      tags:
        - ToolGroups
      summary: Register a tool group
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterToolGroupRequest'
        required: true
  /v1/tools:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListToolsResponse'
      tags:
        - ToolGroups
      summary: List tools with optional tool group
      parameters:
        - name: toolgroup_id
          in: query
          required: false
          schema:
            type: string
  /v1/vector-dbs:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ListVectorDBsResponse'
      tags:
        - VectorDBs
      parameters: []
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/VectorDB'
      tags:
        - VectorDBs
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RegisterVectorDbRequest'
        required: true
  /v1/telemetry/events:
    post:
      responses:
        '200':
          description: OK
      tags:
        - Telemetry
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/LogEventRequest'
        required: true
  /v1/post-training/preference-optimize:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/PostTrainingJob'
      tags:
        - PostTraining (Coming Soon)
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/PreferenceOptimizeRequest'
        required: true
  /v1/tool-runtime/rag-tool/query:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/RAGQueryResult'
      tags:
        - ToolRuntime
      summary: >-
        Query the RAG system for context; typically invoked by the agent
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/QueryRequest'
        required: true
  /v1/vector-io/query:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QueryChunksResponse'
      tags:
        - VectorIO
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/QueryChunksRequest'
        required: true
  /v1/telemetry/spans:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QuerySpansResponse'
      tags:
        - Telemetry
      parameters:
        - name: attribute_filters
          in: query
          required: true
          schema:
            type: array
            items:
              $ref: '#/components/schemas/QueryCondition'
        - name: attributes_to_return
          in: query
          required: true
          schema:
            type: array
            items:
              type: string
        - name: max_depth
          in: query
          required: false
          schema:
            type: integer
  /v1/telemetry/traces:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QueryTracesResponse'
      tags:
        - Telemetry
      parameters:
        - name: attribute_filters
          in: query
          required: false
          schema:
            type: array
            items:
              $ref: '#/components/schemas/QueryCondition'
        - name: limit
          in: query
          required: false
          schema:
            type: integer
        - name: offset
          in: query
          required: false
          schema:
            type: integer
        - name: order_by
          in: query
          required: false
          schema:
            type: array
            items:
              type: string
  /v1/eval/tasks/{task_id}/jobs:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Job'
      tags:
        - Eval
      parameters:
        - name: task_id
          in: path
          required: true
          schema:
            type: string
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RunEvalRequest'
        required: true
  /v1/safety/run-shield:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/RunShieldResponse'
      tags:
        - Safety
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RunShieldRequest'
        required: true
  /v1/telemetry/spans/export:
    post:
      responses:
        '200':
          description: OK
      tags:
        - Telemetry
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SaveSpansToDatasetRequest'
        required: true
  /v1/scoring/score:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ScoreResponse'
      tags:
        - Scoring
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ScoreRequest'
        required: true
  /v1/scoring/score-batch:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ScoreBatchResponse'
      tags:
        - Scoring
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ScoreBatchRequest'
        required: true
  /v1/post-training/supervised-fine-tune:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/PostTrainingJob'
      tags:
        - PostTraining (Coming Soon)
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SupervisedFineTuneRequest'
        required: true
  /v1/synthetic-data-generation/generate:
    post:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SyntheticDataGenerationResponse'
      tags:
        - SyntheticDataGeneration (Coming Soon)
      parameters: []
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SyntheticDataGenerateRequest'
        required: true
  /v1/version:
    get:
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/VersionInfo'
      tags:
        - Inspect
      parameters: []
jsonSchemaDialect: >-
  https://json-schema.org/draft/2020-12/schema
components:
  schemas:
    AppendRowsRequest:
      type: object
      properties:
        dataset_id:
          type: string
        rows:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
      additionalProperties: false
      required:
        - dataset_id
        - rows
    CompletionMessage:
      type: object
      properties:
        role:
          type: string
          const: assistant
          default: assistant
          description: >-
            Must be "assistant" to identify this as the model's response
        content:
          $ref: '#/components/schemas/InterleavedContent'
          description: The content of the model's response
        stop_reason:
          type: string
          enum:
            - end_of_turn
            - end_of_message
            - out_of_tokens
          description: >-
            Reason why the model stopped generating. Options are: - `StopReason.end_of_turn`:
            The model finished generating the entire response. - `StopReason.end_of_message`:
            The model finished generating but generated a partial response -- usually,
            a tool call. The user may call the tool and continue the conversation
            with the tool's response. - `StopReason.out_of_tokens`: The model ran
            out of token budget.
        tool_calls:
          type: array
          items:
            $ref: '#/components/schemas/ToolCall'
          description: >-
            List of tool calls. Each tool call is a ToolCall object.
      additionalProperties: false
      required:
        - role
        - content
        - stop_reason
        - tool_calls
      title: >-
        A message containing the model's (assistant) response in a chat conversation.
    GrammarResponseFormat:
      type: object
      properties:
        type:
          type: string
          const: grammar
          default: grammar
          description: >-
            Must be "grammar" to identify this format type
        bnf:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
          description: >-
            The BNF grammar specification the response should conform to
      additionalProperties: false
      required:
        - type
        - bnf
      title: >-
        Configuration for grammar-guided response generation.
    GreedySamplingStrategy:
      type: object
      properties:
        type:
          type: string
          const: greedy
          default: greedy
      additionalProperties: false
      required:
        - type
    ImageContentItem:
      type: object
      properties:
        type:
          type: string
          const: image
          default: image
          description: >-
            Discriminator type of the content item. Always "image"
        image:
          type: object
          properties:
            url:
              $ref: '#/components/schemas/URL'
              description: >-
                A URL of the image or data URL in the format of data:image/{type};base64,{data}.
                Note that URL could have length limits.
            data:
              type: string
              contentEncoding: base64
              description: base64 encoded image data as string
          additionalProperties: false
          description: >-
            Image as a base64 encoded string or an URL
      additionalProperties: false
      required:
        - type
        - image
      title: A image content item
    InterleavedContent:
      oneOf:
        - type: string
        - $ref: '#/components/schemas/InterleavedContentItem'
        - type: array
          items:
            $ref: '#/components/schemas/InterleavedContentItem'
    InterleavedContentItem:
      oneOf:
        - $ref: '#/components/schemas/ImageContentItem'
        - $ref: '#/components/schemas/TextContentItem'
      discriminator:
        propertyName: type
        mapping:
          image: '#/components/schemas/ImageContentItem'
          text: '#/components/schemas/TextContentItem'
    JsonSchemaResponseFormat:
      type: object
      properties:
        type:
          type: string
          const: json_schema
          default: json_schema
          description: >-
            Must be "json_schema" to identify this format type
        json_schema:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
          description: >-
            The JSON schema the response should conform to. In a Python SDK, this
            is often a `pydantic` model.
      additionalProperties: false
      required:
        - type
        - json_schema
      title: >-
        Configuration for JSON schema-guided response generation.
    Message:
      oneOf:
        - $ref: '#/components/schemas/UserMessage'
        - $ref: '#/components/schemas/SystemMessage'
        - $ref: '#/components/schemas/ToolResponseMessage'
        - $ref: '#/components/schemas/CompletionMessage'
      discriminator:
        propertyName: role
        mapping:
          user: '#/components/schemas/UserMessage'
          system: '#/components/schemas/SystemMessage'
          tool: '#/components/schemas/ToolResponseMessage'
          assistant: '#/components/schemas/CompletionMessage'
    ResponseFormat:
      oneOf:
        - $ref: '#/components/schemas/JsonSchemaResponseFormat'
        - $ref: '#/components/schemas/GrammarResponseFormat'
      discriminator:
        propertyName: type
        mapping:
          json_schema: '#/components/schemas/JsonSchemaResponseFormat'
          grammar: '#/components/schemas/GrammarResponseFormat'
    SamplingParams:
      type: object
      properties:
        strategy:
          $ref: '#/components/schemas/SamplingStrategy'
        max_tokens:
          type: integer
          default: 0
        repetition_penalty:
          type: number
          default: 1.0
      additionalProperties: false
      required:
        - strategy
    SamplingStrategy:
      oneOf:
        - $ref: '#/components/schemas/GreedySamplingStrategy'
        - $ref: '#/components/schemas/TopPSamplingStrategy'
        - $ref: '#/components/schemas/TopKSamplingStrategy'
      discriminator:
        propertyName: type
        mapping:
          greedy: '#/components/schemas/GreedySamplingStrategy'
          top_p: '#/components/schemas/TopPSamplingStrategy'
          top_k: '#/components/schemas/TopKSamplingStrategy'
    SystemMessage:
      type: object
      properties:
        role:
          type: string
          const: system
          default: system
          description: >-
            Must be "system" to identify this as a system message
        content:
          $ref: '#/components/schemas/InterleavedContent'
          description: >-
            The content of the "system prompt". If multiple system messages are provided,
            they are concatenated. The underlying Llama Stack code may also add other
            system messages (for example, for formatting tool definitions).
      additionalProperties: false
      required:
        - role
        - content
      title: >-
        A system message providing instructions or context to the model.
    TextContentItem:
      type: object
      properties:
        type:
          type: string
          const: text
          default: text
          description: >-
            Discriminator type of the content item. Always "text"
        text:
          type: string
          description: Text content
      additionalProperties: false
      required:
        - type
        - text
      title: A text content item
    ToolCall:
      type: object
      properties:
        call_id:
          type: string
        tool_name:
          oneOf:
            - type: string
              enum:
                - brave_search
                - wolfram_alpha
                - photogen
                - code_interpreter
            - type: string
        arguments:
          type: object
          additionalProperties:
            oneOf:
              - type: string
              - type: integer
              - type: number
              - type: boolean
              - type: 'null'
              - type: array
                items:
                  oneOf:
                    - type: string
                    - type: integer
                    - type: number
                    - type: boolean
                    - type: 'null'
              - type: object
                additionalProperties:
                  oneOf:
                    - type: string
                    - type: integer
                    - type: number
                    - type: boolean
                    - type: 'null'
      additionalProperties: false
      required:
        - call_id
        - tool_name
        - arguments
    ToolDefinition:
      type: object
      properties:
        tool_name:
          oneOf:
            - type: string
              enum:
                - brave_search
                - wolfram_alpha
                - photogen
                - code_interpreter
            - type: string
        description:
          type: string
        parameters:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ToolParamDefinition'
      additionalProperties: false
      required:
        - tool_name
    ToolParamDefinition:
      type: object
      properties:
        param_type:
          type: string
        description:
          type: string
        required:
          type: boolean
          default: true
        default:
          oneOf:
            - type: 'null'
            - type: boolean
            - type: number
            - type: string
            - type: array
            - type: object
      additionalProperties: false
      required:
        - param_type
    ToolResponseMessage:
      type: object
      properties:
        role:
          type: string
          const: tool
          default: tool
          description: >-
            Must be "tool" to identify this as a tool response
        call_id:
          type: string
          description: >-
            Unique identifier for the tool call this response is for
        tool_name:
          oneOf:
            - type: string
              enum:
                - brave_search
                - wolfram_alpha
                - photogen
                - code_interpreter
            - type: string
          description: Name of the tool that was called
        content:
          $ref: '#/components/schemas/InterleavedContent'
          description: The response content from the tool
      additionalProperties: false
      required:
        - role
        - call_id
        - tool_name
        - content
      title: >-
        A message representing the result of a tool invocation.
    TopKSamplingStrategy:
      type: object
      properties:
        type:
          type: string
          const: top_k
          default: top_k
        top_k:
          type: integer
      additionalProperties: false
      required:
        - type
        - top_k
    TopPSamplingStrategy:
      type: object
      properties:
        type:
          type: string
          const: top_p
          default: top_p
        temperature:
          type: number
        top_p:
          type: number
          default: 0.95
      additionalProperties: false
      required:
        - type
    URL:
      type: object
      properties:
        uri:
          type: string
      additionalProperties: false
      required:
        - uri
    UserMessage:
      type: object
      properties:
        role:
          type: string
          const: user
          default: user
          description: >-
            Must be "user" to identify this as a user message
        content:
          $ref: '#/components/schemas/InterleavedContent'
          description: >-
            The content of the message, which can include text and other media
        context:
          $ref: '#/components/schemas/InterleavedContent'
          description: >-
            (Optional) This field is used internally by Llama Stack to pass RAG context.
            This field may be removed in the API in the future.
      additionalProperties: false
      required:
        - role
        - content
      title: >-
        A message from the user in a chat conversation.
    BatchChatCompletionRequest:
      type: object
      properties:
        model:
          type: string
        messages_batch:
          type: array
          items:
            type: array
            items:
              $ref: '#/components/schemas/Message'
        sampling_params:
          $ref: '#/components/schemas/SamplingParams'
        tools:
          type: array
          items:
            $ref: '#/components/schemas/ToolDefinition'
        tool_choice:
          type: string
          enum:
            - auto
            - required
          title: >-
            Whether tool use is required or automatic. This is a hint to the model
            which may not be followed. It depends on the Instruction Following capabilities
            of the model.
        tool_prompt_format:
          type: string
          enum:
            - json
            - function_tag
            - python_list
          title: >-
            Prompt format for calling custom / zero shot tools.
        response_format:
          $ref: '#/components/schemas/ResponseFormat'
        logprobs:
          type: object
          properties:
            top_k:
              type: integer
              default: 0
              description: >-
                How many tokens (for each position) to return log probabilities for.
          additionalProperties: false
      additionalProperties: false
      required:
        - model
        - messages_batch
    BatchChatCompletionResponse:
      type: object
      properties:
        batch:
          type: array
          items:
            $ref: '#/components/schemas/ChatCompletionResponse'
      additionalProperties: false
      required:
        - batch
    ChatCompletionResponse:
      type: object
      properties:
        completion_message:
          $ref: '#/components/schemas/CompletionMessage'
          description: The complete response message
        logprobs:
          type: array
          items:
            $ref: '#/components/schemas/TokenLogProbs'
          description: >-
            Optional log probabilities for generated tokens
      additionalProperties: false
      required:
        - completion_message
      title: Response from a chat completion request.
    TokenLogProbs:
      type: object
      properties:
        logprobs_by_token:
          type: object
          additionalProperties:
            type: number
          description: >-
            Dictionary mapping tokens to their log probabilities
      additionalProperties: false
      required:
        - logprobs_by_token
      title: Log probabilities for generated tokens.
    BatchCompletionRequest:
      type: object
      properties:
        model:
          type: string
        content_batch:
          type: array
          items:
            $ref: '#/components/schemas/InterleavedContent'
        sampling_params:
          $ref: '#/components/schemas/SamplingParams'
        response_format:
          $ref: '#/components/schemas/ResponseFormat'
        logprobs:
          type: object
          properties:
            top_k:
              type: integer
              default: 0
              description: >-
                How many tokens (for each position) to return log probabilities for.
          additionalProperties: false
      additionalProperties: false
      required:
        - model
        - content_batch
    BatchCompletionResponse:
      type: object
      properties:
        batch:
          type: array
          items:
            $ref: '#/components/schemas/CompletionResponse'
      additionalProperties: false
      required:
        - batch
    CompletionResponse:
      type: object
      properties:
        content:
          type: string
          description: The generated completion text
        stop_reason:
          type: string
          enum:
            - end_of_turn
            - end_of_message
            - out_of_tokens
          description: Reason why generation stopped
        logprobs:
          type: array
          items:
            $ref: '#/components/schemas/TokenLogProbs'
          description: >-
            Optional log probabilities for generated tokens
      additionalProperties: false
      required:
        - content
        - stop_reason
      title: Response from a completion request.
    CancelTrainingJobRequest:
      type: object
      properties:
        job_uuid:
          type: string
      additionalProperties: false
      required:
        - job_uuid
    ChatCompletionRequest:
      type: object
      properties:
        model_id:
          type: string
          description: >-
            The identifier of the model to use. The model must be registered with
            Llama Stack and available via the /models endpoint.
        messages:
          type: array
          items:
            $ref: '#/components/schemas/Message'
          description: List of messages in the conversation
        sampling_params:
          $ref: '#/components/schemas/SamplingParams'
          description: >-
            Parameters to control the sampling strategy
        tools:
          type: array
          items:
            $ref: '#/components/schemas/ToolDefinition'
          description: >-
            (Optional) List of tool definitions available to the model
        tool_choice:
          type: string
          enum:
            - auto
            - required
          description: >-
            (Optional) Whether tool use is required or automatic. Defaults to ToolChoice.auto.
        tool_prompt_format:
          type: string
          enum:
            - json
            - function_tag
            - python_list
          description: >-
            (Optional) Instructs the model how to format tool calls. By default, Llama
            Stack will attempt to use a format that is best adapted to the model.
            - `ToolPromptFormat.json`: The tool calls are formatted as a JSON object.
            - `ToolPromptFormat.function_tag`: The tool calls are enclosed in a <function=function_name>
            tag. - `ToolPromptFormat.python_list`: The tool calls are output as Python
            syntax -- a list of function calls.
        response_format:
          $ref: '#/components/schemas/ResponseFormat'
          description: >-
            (Optional) Grammar specification for guided (structured) decoding. There
            are two options: - `ResponseFormat.json_schema`: The grammar is a JSON
            schema. Most providers support this format. - `ResponseFormat.grammar`:
            The grammar is a BNF grammar. This format is more flexible, but not all
            providers support it.
        stream:
          type: boolean
          description: >-
            (Optional) If True, generate an SSE event stream of the response. Defaults
            to False.
        logprobs:
          type: object
          properties:
            top_k:
              type: integer
              default: 0
              description: >-
                How many tokens (for each position) to return log probabilities for.
          additionalProperties: false
          description: >-
            (Optional) If specified, log probabilities for each token position will
            be returned.
      additionalProperties: false
      required:
        - model_id
        - messages
    ChatCompletionResponseEvent:
      type: object
      properties:
        event_type:
          type: string
          enum:
            - start
            - complete
            - progress
          description: Type of the event
        delta:
          $ref: '#/components/schemas/ContentDelta'
          description: >-
            Content generated since last event. This can be one or more tokens, or
            a tool call.
        logprobs:
          type: array
          items:
            $ref: '#/components/schemas/TokenLogProbs'
          description: >-
            Optional log probabilities for generated tokens
        stop_reason:
          type: string
          enum:
            - end_of_turn
            - end_of_message
            - out_of_tokens
          description: >-
            Optional reason why generation stopped, if complete
      additionalProperties: false
      required:
        - event_type
        - delta
      title: >-
        An event during chat completion generation.
    ChatCompletionResponseStreamChunk:
      type: object
      properties:
        event:
          $ref: '#/components/schemas/ChatCompletionResponseEvent'
          description: The event containing the new content
      additionalProperties: false
      required:
        - event
      title: >-
        A chunk of a streamed chat completion response.
    ContentDelta:
      oneOf:
        - $ref: '#/components/schemas/TextDelta'
        - $ref: '#/components/schemas/ImageDelta'
        - $ref: '#/components/schemas/ToolCallDelta'
      discriminator:
        propertyName: type
        mapping:
          text: '#/components/schemas/TextDelta'
          image: '#/components/schemas/ImageDelta'
          tool_call: '#/components/schemas/ToolCallDelta'
    ImageDelta:
      type: object
      properties:
        type:
          type: string
          const: image
          default: image
        image:
          type: string
          contentEncoding: base64
      additionalProperties: false
      required:
        - type
        - image
    TextDelta:
      type: object
      properties:
        type:
          type: string
          const: text
          default: text
        text:
          type: string
      additionalProperties: false
      required:
        - type
        - text
    ToolCallDelta:
      type: object
      properties:
        type:
          type: string
          const: tool_call
          default: tool_call
        tool_call:
          oneOf:
            - type: string
            - $ref: '#/components/schemas/ToolCall'
        parse_status:
          type: string
          enum:
            - started
            - in_progress
            - failed
            - succeeded
      additionalProperties: false
      required:
        - type
        - tool_call
        - parse_status
    CompletionRequest:
      type: object
      properties:
        model_id:
          type: string
          description: >-
            The identifier of the model to use. The model must be registered with
            Llama Stack and available via the /models endpoint.
        content:
          $ref: '#/components/schemas/InterleavedContent'
          description: The content to generate a completion for
        sampling_params:
          $ref: '#/components/schemas/SamplingParams'
          description: >-
            (Optional) Parameters to control the sampling strategy
        response_format:
          $ref: '#/components/schemas/ResponseFormat'
          description: >-
            (Optional) Grammar specification for guided (structured) decoding
        stream:
          type: boolean
          description: >-
            (Optional) If True, generate an SSE event stream of the response. Defaults
            to False.
        logprobs:
          type: object
          properties:
            top_k:
              type: integer
              default: 0
              description: >-
                How many tokens (for each position) to return log probabilities for.
          additionalProperties: false
          description: >-
            (Optional) If specified, log probabilities for each token position will
            be returned.
      additionalProperties: false
      required:
        - model_id
        - content
    CompletionResponseStreamChunk:
      type: object
      properties:
        delta:
          type: string
          description: >-
            New content generated since last chunk. This can be one or more tokens.
        stop_reason:
          type: string
          enum:
            - end_of_turn
            - end_of_message
            - out_of_tokens
          description: >-
            Optional reason why generation stopped, if complete
        logprobs:
          type: array
          items:
            $ref: '#/components/schemas/TokenLogProbs'
          description: >-
            Optional log probabilities for generated tokens
      additionalProperties: false
      required:
        - delta
      title: >-
        A chunk of a streamed completion response.
    AgentConfig:
      type: object
      properties:
        sampling_params:
          $ref: '#/components/schemas/SamplingParams'
        input_shields:
          type: array
          items:
            type: string
        output_shields:
          type: array
          items:
            type: string
        toolgroups:
          type: array
          items:
            $ref: '#/components/schemas/AgentTool'
        client_tools:
          type: array
          items:
            $ref: '#/components/schemas/ToolDef'
        tool_choice:
          type: string
          enum:
            - auto
            - required
          title: >-
            Whether tool use is required or automatic. This is a hint to the model
            which may not be followed. It depends on the Instruction Following capabilities
            of the model.
          default: auto
        tool_prompt_format:
          type: string
          enum:
            - json
            - function_tag
            - python_list
          title: >-
            Prompt format for calling custom / zero shot tools.
        max_infer_iters:
          type: integer
          default: 10
        model:
          type: string
        instructions:
          type: string
        enable_session_persistence:
          type: boolean
        response_format:
          $ref: '#/components/schemas/ResponseFormat'
      additionalProperties: false
      required:
        - max_infer_iters
        - model
        - instructions
        - enable_session_persistence
    AgentTool:
      oneOf:
        - type: string
        - type: object
          properties:
            name:
              type: string
            args:
              type: object
              additionalProperties:
                oneOf:
                  - type: 'null'
                  - type: boolean
                  - type: number
                  - type: string
                  - type: array
                  - type: object
          additionalProperties: false
          required:
            - name
            - args
    ToolDef:
      type: object
      properties:
        name:
          type: string
        description:
          type: string
        parameters:
          type: array
          items:
            $ref: '#/components/schemas/ToolParameter'
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - name
    ToolParameter:
      type: object
      properties:
        name:
          type: string
        parameter_type:
          type: string
        description:
          type: string
        required:
          type: boolean
          default: true
        default:
          oneOf:
            - type: 'null'
            - type: boolean
            - type: number
            - type: string
            - type: array
            - type: object
      additionalProperties: false
      required:
        - name
        - parameter_type
        - description
        - required
    CreateAgentRequest:
      type: object
      properties:
        agent_config:
          $ref: '#/components/schemas/AgentConfig'
      additionalProperties: false
      required:
        - agent_config
    AgentCreateResponse:
      type: object
      properties:
        agent_id:
          type: string
      additionalProperties: false
      required:
        - agent_id
    CreateAgentSessionRequest:
      type: object
      properties:
        session_name:
          type: string
      additionalProperties: false
      required:
        - session_name
    AgentSessionCreateResponse:
      type: object
      properties:
        session_id:
          type: string
      additionalProperties: false
      required:
        - session_id
    CreateAgentTurnRequest:
      type: object
      properties:
        messages:
          type: array
          items:
            oneOf:
              - $ref: '#/components/schemas/UserMessage'
              - $ref: '#/components/schemas/ToolResponseMessage'
        stream:
          type: boolean
        documents:
          type: array
          items:
            type: object
            properties:
              content:
                oneOf:
                  - type: string
                  - $ref: '#/components/schemas/InterleavedContentItem'
                  - type: array
                    items:
                      $ref: '#/components/schemas/InterleavedContentItem'
                  - $ref: '#/components/schemas/URL'
              mime_type:
                type: string
            additionalProperties: false
            required:
              - content
              - mime_type
        toolgroups:
          type: array
          items:
            $ref: '#/components/schemas/AgentTool'
      additionalProperties: false
      required:
        - messages
    InferenceStep:
      type: object
      properties:
        turn_id:
          type: string
        step_id:
          type: string
        started_at:
          type: string
          format: date-time
        completed_at:
          type: string
          format: date-time
        step_type:
          type: string
          const: inference
          default: inference
        model_response:
          $ref: '#/components/schemas/CompletionMessage'
      additionalProperties: false
      required:
        - turn_id
        - step_id
        - step_type
        - model_response
    MemoryRetrievalStep:
      type: object
      properties:
        turn_id:
          type: string
        step_id:
          type: string
        started_at:
          type: string
          format: date-time
        completed_at:
          type: string
          format: date-time
        step_type:
          type: string
          const: memory_retrieval
          default: memory_retrieval
        vector_db_ids:
          type: string
        inserted_context:
          $ref: '#/components/schemas/InterleavedContent'
      additionalProperties: false
      required:
        - turn_id
        - step_id
        - step_type
        - vector_db_ids
        - inserted_context
    SafetyViolation:
      type: object
      properties:
        violation_level:
          $ref: '#/components/schemas/ViolationLevel'
        user_message:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - violation_level
        - metadata
    ShieldCallStep:
      type: object
      properties:
        turn_id:
          type: string
        step_id:
          type: string
        started_at:
          type: string
          format: date-time
        completed_at:
          type: string
          format: date-time
        step_type:
          type: string
          const: shield_call
          default: shield_call
        violation:
          $ref: '#/components/schemas/SafetyViolation'
      additionalProperties: false
      required:
        - turn_id
        - step_id
        - step_type
    ToolExecutionStep:
      type: object
      properties:
        turn_id:
          type: string
        step_id:
          type: string
        started_at:
          type: string
          format: date-time
        completed_at:
          type: string
          format: date-time
        step_type:
          type: string
          const: tool_execution
          default: tool_execution
        tool_calls:
          type: array
          items:
            $ref: '#/components/schemas/ToolCall'
        tool_responses:
          type: array
          items:
            $ref: '#/components/schemas/ToolResponse'
      additionalProperties: false
      required:
        - turn_id
        - step_id
        - step_type
        - tool_calls
        - tool_responses
    ToolResponse:
      type: object
      properties:
        call_id:
          type: string
        tool_name:
          oneOf:
            - type: string
              enum:
                - brave_search
                - wolfram_alpha
                - photogen
                - code_interpreter
            - type: string
        content:
          $ref: '#/components/schemas/InterleavedContent'
      additionalProperties: false
      required:
        - call_id
        - tool_name
        - content
    Turn:
      type: object
      properties:
        turn_id:
          type: string
        session_id:
          type: string
        input_messages:
          type: array
          items:
            oneOf:
              - $ref: '#/components/schemas/UserMessage'
              - $ref: '#/components/schemas/ToolResponseMessage'
        steps:
          type: array
          items:
            oneOf:
              - $ref: '#/components/schemas/InferenceStep'
              - $ref: '#/components/schemas/ToolExecutionStep'
              - $ref: '#/components/schemas/ShieldCallStep'
              - $ref: '#/components/schemas/MemoryRetrievalStep'
            discriminator:
              propertyName: step_type
              mapping:
                inference: '#/components/schemas/InferenceStep'
                tool_execution: '#/components/schemas/ToolExecutionStep'
                shield_call: '#/components/schemas/ShieldCallStep'
                memory_retrieval: '#/components/schemas/MemoryRetrievalStep'
        output_message:
          $ref: '#/components/schemas/CompletionMessage'
        output_attachments:
          type: array
          items:
            type: object
            properties:
              content:
                oneOf:
                  - type: string
                  - $ref: '#/components/schemas/InterleavedContentItem'
                  - type: array
                    items:
                      $ref: '#/components/schemas/InterleavedContentItem'
                  - $ref: '#/components/schemas/URL'
              mime_type:
                type: string
            additionalProperties: false
            required:
              - content
              - mime_type
        started_at:
          type: string
          format: date-time
        completed_at:
          type: string
          format: date-time
      additionalProperties: false
      required:
        - turn_id
        - session_id
        - input_messages
        - steps
        - output_message
        - output_attachments
        - started_at
      title: >-
        A single turn in an interaction with an Agentic System.
    ViolationLevel:
      type: string
      enum:
        - info
        - warn
        - error
    AgentTurnResponseEvent:
      type: object
      properties:
        payload:
          $ref: '#/components/schemas/AgentTurnResponseEventPayload'
      additionalProperties: false
      required:
        - payload
    AgentTurnResponseEventPayload:
      oneOf:
        - $ref: '#/components/schemas/AgentTurnResponseStepStartPayload'
        - $ref: '#/components/schemas/AgentTurnResponseStepProgressPayload'
        - $ref: '#/components/schemas/AgentTurnResponseStepCompletePayload'
        - $ref: '#/components/schemas/AgentTurnResponseTurnStartPayload'
        - $ref: '#/components/schemas/AgentTurnResponseTurnCompletePayload'
      discriminator:
        propertyName: event_type
        mapping:
          step_start: '#/components/schemas/AgentTurnResponseStepStartPayload'
          step_progress: '#/components/schemas/AgentTurnResponseStepProgressPayload'
          step_complete: '#/components/schemas/AgentTurnResponseStepCompletePayload'
          turn_start: '#/components/schemas/AgentTurnResponseTurnStartPayload'
          turn_complete: '#/components/schemas/AgentTurnResponseTurnCompletePayload'
    AgentTurnResponseStepCompletePayload:
      type: object
      properties:
        event_type:
          type: string
          const: step_complete
          default: step_complete
        step_type:
          type: string
          enum:
            - inference
            - tool_execution
            - shield_call
            - memory_retrieval
        step_id:
          type: string
        step_details:
          oneOf:
            - $ref: '#/components/schemas/InferenceStep'
            - $ref: '#/components/schemas/ToolExecutionStep'
            - $ref: '#/components/schemas/ShieldCallStep'
            - $ref: '#/components/schemas/MemoryRetrievalStep'
          discriminator:
            propertyName: step_type
            mapping:
              inference: '#/components/schemas/InferenceStep'
              tool_execution: '#/components/schemas/ToolExecutionStep'
              shield_call: '#/components/schemas/ShieldCallStep'
              memory_retrieval: '#/components/schemas/MemoryRetrievalStep'
      additionalProperties: false
      required:
        - event_type
        - step_type
        - step_id
        - step_details
    AgentTurnResponseStepProgressPayload:
      type: object
      properties:
        event_type:
          type: string
          const: step_progress
          default: step_progress
        step_type:
          type: string
          enum:
            - inference
            - tool_execution
            - shield_call
            - memory_retrieval
        step_id:
          type: string
        delta:
          $ref: '#/components/schemas/ContentDelta'
      additionalProperties: false
      required:
        - event_type
        - step_type
        - step_id
        - delta
    AgentTurnResponseStepStartPayload:
      type: object
      properties:
        event_type:
          type: string
          const: step_start
          default: step_start
        step_type:
          type: string
          enum:
            - inference
            - tool_execution
            - shield_call
            - memory_retrieval
        step_id:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - event_type
        - step_type
        - step_id
    AgentTurnResponseStreamChunk:
      type: object
      properties:
        event:
          $ref: '#/components/schemas/AgentTurnResponseEvent'
      additionalProperties: false
      required:
        - event
      title: streamed agent turn completion response.
    AgentTurnResponseTurnCompletePayload:
      type: object
      properties:
        event_type:
          type: string
          const: turn_complete
          default: turn_complete
        turn:
          $ref: '#/components/schemas/Turn'
      additionalProperties: false
      required:
        - event_type
        - turn
    AgentTurnResponseTurnStartPayload:
      type: object
      properties:
        event_type:
          type: string
          const: turn_start
          default: turn_start
        turn_id:
          type: string
      additionalProperties: false
      required:
        - event_type
        - turn_id
    EmbeddingsRequest:
      type: object
      properties:
        model_id:
          type: string
          description: >-
            The identifier of the model to use. The model must be an embedding model
            registered with Llama Stack and available via the /models endpoint.
        contents:
          type: array
          items:
            $ref: '#/components/schemas/InterleavedContent'
          description: >-
            List of contents to generate embeddings for. Note that content can be
            multimodal. The behavior depends on the model and provider. Some models
            may only support text.
      additionalProperties: false
      required:
        - model_id
        - contents
    EmbeddingsResponse:
      type: object
      properties:
        embeddings:
          type: array
          items:
            type: array
            items:
              type: number
          description: >-
            List of embedding vectors, one per input content. Each embedding is a
            list of floats. The dimensionality of the embedding is model-specific;
            you can check model metadata using /models/{model_id}
      additionalProperties: false
      required:
        - embeddings
      title: >-
        Response containing generated embeddings.
    AgentCandidate:
      type: object
      properties:
        type:
          type: string
          const: agent
          default: agent
        config:
          $ref: '#/components/schemas/AgentConfig'
      additionalProperties: false
      required:
        - type
        - config
    AggregationFunctionType:
      type: string
      enum:
        - average
        - median
        - categorical_count
        - accuracy
    AppEvalTaskConfig:
      type: object
      properties:
        type:
          type: string
          const: app
          default: app
        eval_candidate:
          $ref: '#/components/schemas/EvalCandidate'
        scoring_params:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ScoringFnParams'
        num_examples:
          type: integer
      additionalProperties: false
      required:
        - type
        - eval_candidate
        - scoring_params
    BasicScoringFnParams:
      type: object
      properties:
        type:
          type: string
          const: basic
          default: basic
        aggregation_functions:
          type: array
          items:
            $ref: '#/components/schemas/AggregationFunctionType'
      additionalProperties: false
      required:
        - type
    BenchmarkEvalTaskConfig:
      type: object
      properties:
        type:
          type: string
          const: benchmark
          default: benchmark
        eval_candidate:
          $ref: '#/components/schemas/EvalCandidate'
        num_examples:
          type: integer
      additionalProperties: false
      required:
        - type
        - eval_candidate
    EvalCandidate:
      oneOf:
        - $ref: '#/components/schemas/ModelCandidate'
        - $ref: '#/components/schemas/AgentCandidate'
      discriminator:
        propertyName: type
        mapping:
          model: '#/components/schemas/ModelCandidate'
          agent: '#/components/schemas/AgentCandidate'
    EvalTaskConfig:
      oneOf:
        - $ref: '#/components/schemas/BenchmarkEvalTaskConfig'
        - $ref: '#/components/schemas/AppEvalTaskConfig'
      discriminator:
        propertyName: type
        mapping:
          benchmark: '#/components/schemas/BenchmarkEvalTaskConfig'
          app: '#/components/schemas/AppEvalTaskConfig'
    LLMAsJudgeScoringFnParams:
      type: object
      properties:
        type:
          type: string
          const: llm_as_judge
          default: llm_as_judge
        judge_model:
          type: string
        prompt_template:
          type: string
        judge_score_regexes:
          type: array
          items:
            type: string
        aggregation_functions:
          type: array
          items:
            $ref: '#/components/schemas/AggregationFunctionType'
      additionalProperties: false
      required:
        - type
        - judge_model
    ModelCandidate:
      type: object
      properties:
        type:
          type: string
          const: model
          default: model
        model:
          type: string
        sampling_params:
          $ref: '#/components/schemas/SamplingParams'
        system_message:
          $ref: '#/components/schemas/SystemMessage'
      additionalProperties: false
      required:
        - type
        - model
        - sampling_params
    RegexParserScoringFnParams:
      type: object
      properties:
        type:
          type: string
          const: regex_parser
          default: regex_parser
        parsing_regexes:
          type: array
          items:
            type: string
        aggregation_functions:
          type: array
          items:
            $ref: '#/components/schemas/AggregationFunctionType'
      additionalProperties: false
      required:
        - type
    ScoringFnParams:
      oneOf:
        - $ref: '#/components/schemas/LLMAsJudgeScoringFnParams'
        - $ref: '#/components/schemas/RegexParserScoringFnParams'
        - $ref: '#/components/schemas/BasicScoringFnParams'
      discriminator:
        propertyName: type
        mapping:
          llm_as_judge: '#/components/schemas/LLMAsJudgeScoringFnParams'
          regex_parser: '#/components/schemas/RegexParserScoringFnParams'
          basic: '#/components/schemas/BasicScoringFnParams'
    EvaluateRowsRequest:
      type: object
      properties:
        input_rows:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
        scoring_functions:
          type: array
          items:
            type: string
        task_config:
          $ref: '#/components/schemas/EvalTaskConfig'
      additionalProperties: false
      required:
        - input_rows
        - scoring_functions
        - task_config
    EvaluateResponse:
      type: object
      properties:
        generations:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
        scores:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ScoringResult'
      additionalProperties: false
      required:
        - generations
        - scores
    ScoringResult:
      type: object
      properties:
        score_rows:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
        aggregated_results:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - score_rows
        - aggregated_results
    Session:
      type: object
      properties:
        session_id:
          type: string
        session_name:
          type: string
        turns:
          type: array
          items:
            $ref: '#/components/schemas/Turn'
        started_at:
          type: string
          format: date-time
      additionalProperties: false
      required:
        - session_id
        - session_name
        - turns
        - started_at
      title: >-
        A single session of an interaction with an Agentic System.
    AgentStepResponse:
      type: object
      properties:
        step:
          oneOf:
            - $ref: '#/components/schemas/InferenceStep'
            - $ref: '#/components/schemas/ToolExecutionStep'
            - $ref: '#/components/schemas/ShieldCallStep'
            - $ref: '#/components/schemas/MemoryRetrievalStep'
          discriminator:
            propertyName: step_type
            mapping:
              inference: '#/components/schemas/InferenceStep'
              tool_execution: '#/components/schemas/ToolExecutionStep'
              shield_call: '#/components/schemas/ShieldCallStep'
              memory_retrieval: '#/components/schemas/MemoryRetrievalStep'
      additionalProperties: false
      required:
        - step
    AgentTurnInputType:
      type: object
      properties:
        type:
          type: string
          const: agent_turn_input
          default: agent_turn_input
      additionalProperties: false
      required:
        - type
    ArrayType:
      type: object
      properties:
        type:
          type: string
          const: array
          default: array
      additionalProperties: false
      required:
        - type
    BooleanType:
      type: object
      properties:
        type:
          type: string
          const: boolean
          default: boolean
      additionalProperties: false
      required:
        - type
    ChatCompletionInputType:
      type: object
      properties:
        type:
          type: string
          const: chat_completion_input
          default: chat_completion_input
      additionalProperties: false
      required:
        - type
    CompletionInputType:
      type: object
      properties:
        type:
          type: string
          const: completion_input
          default: completion_input
      additionalProperties: false
      required:
        - type
    Dataset:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: dataset
          default: dataset
        dataset_schema:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ParamType'
        url:
          $ref: '#/components/schemas/URL'
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
        - dataset_schema
        - url
        - metadata
    JsonType:
      type: object
      properties:
        type:
          type: string
          const: json
          default: json
      additionalProperties: false
      required:
        - type
    NumberType:
      type: object
      properties:
        type:
          type: string
          const: number
          default: number
      additionalProperties: false
      required:
        - type
    ObjectType:
      type: object
      properties:
        type:
          type: string
          const: object
          default: object
      additionalProperties: false
      required:
        - type
    ParamType:
      oneOf:
        - $ref: '#/components/schemas/StringType'
        - $ref: '#/components/schemas/NumberType'
        - $ref: '#/components/schemas/BooleanType'
        - $ref: '#/components/schemas/ArrayType'
        - $ref: '#/components/schemas/ObjectType'
        - $ref: '#/components/schemas/JsonType'
        - $ref: '#/components/schemas/UnionType'
        - $ref: '#/components/schemas/ChatCompletionInputType'
        - $ref: '#/components/schemas/CompletionInputType'
        - $ref: '#/components/schemas/AgentTurnInputType'
      discriminator:
        propertyName: type
        mapping:
          string: '#/components/schemas/StringType'
          number: '#/components/schemas/NumberType'
          boolean: '#/components/schemas/BooleanType'
          array: '#/components/schemas/ArrayType'
          object: '#/components/schemas/ObjectType'
          json: '#/components/schemas/JsonType'
          union: '#/components/schemas/UnionType'
          chat_completion_input: '#/components/schemas/ChatCompletionInputType'
          completion_input: '#/components/schemas/CompletionInputType'
          agent_turn_input: '#/components/schemas/AgentTurnInputType'
    StringType:
      type: object
      properties:
        type:
          type: string
          const: string
          default: string
      additionalProperties: false
      required:
        - type
    UnionType:
      type: object
      properties:
        type:
          type: string
          const: union
          default: union
      additionalProperties: false
      required:
        - type
    EvalTask:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: eval_task
          default: eval_task
        dataset_id:
          type: string
        scoring_functions:
          type: array
          items:
            type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
        - dataset_id
        - scoring_functions
        - metadata
    Model:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: model
          default: model
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        model_type:
          $ref: '#/components/schemas/ModelType'
          default: llm
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
        - metadata
        - model_type
    ModelType:
      type: string
      enum:
        - llm
        - embedding
    PaginatedRowsResult:
      type: object
      properties:
        rows:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
        total_count:
          type: integer
        next_page_token:
          type: string
      additionalProperties: false
      required:
        - rows
        - total_count
    ScoringFn:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: scoring_function
          default: scoring_function
        description:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        return_type:
          $ref: '#/components/schemas/ParamType'
        params:
          $ref: '#/components/schemas/ScoringFnParams'
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
        - metadata
        - return_type
    Shield:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: shield
          default: shield
        params:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
      title: >-
        A safety shield resource that can be used to check content
    Span:
      type: object
      properties:
        span_id:
          type: string
        trace_id:
          type: string
        parent_span_id:
          type: string
        name:
          type: string
        start_time:
          type: string
          format: date-time
        end_time:
          type: string
          format: date-time
        attributes:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - span_id
        - trace_id
        - name
        - start_time
    SpanStatus:
      type: string
      enum:
        - ok
        - error
    SpanWithStatus:
      type: object
      properties:
        span_id:
          type: string
        trace_id:
          type: string
        parent_span_id:
          type: string
        name:
          type: string
        start_time:
          type: string
          format: date-time
        end_time:
          type: string
          format: date-time
        attributes:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        status:
          $ref: '#/components/schemas/SpanStatus'
      additionalProperties: false
      required:
        - span_id
        - trace_id
        - name
        - start_time
    QuerySpanTreeResponse:
      type: object
      properties:
        data:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/SpanWithStatus'
      additionalProperties: false
      required:
        - data
    Tool:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: tool
          default: tool
        toolgroup_id:
          type: string
        tool_host:
          $ref: '#/components/schemas/ToolHost'
        description:
          type: string
        parameters:
          type: array
          items:
            $ref: '#/components/schemas/ToolParameter'
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
        - toolgroup_id
        - tool_host
        - description
        - parameters
    ToolHost:
      type: string
      enum:
        - distribution
        - client
        - model_context_protocol
    ToolGroup:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: tool_group
          default: tool_group
        mcp_endpoint:
          $ref: '#/components/schemas/URL'
        args:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
    Trace:
      type: object
      properties:
        trace_id:
          type: string
        root_span_id:
          type: string
        start_time:
          type: string
          format: date-time
        end_time:
          type: string
          format: date-time
      additionalProperties: false
      required:
        - trace_id
        - root_span_id
        - start_time
    Checkpoint:
      description: Checkpoint created during training runs
    PostTrainingJobArtifactsResponse:
      type: object
      properties:
        job_uuid:
          type: string
        checkpoints:
          type: array
          items:
            $ref: '#/components/schemas/Checkpoint'
      additionalProperties: false
      required:
        - job_uuid
        - checkpoints
      title: Artifacts of a finetuning job.
    JobStatus:
      type: string
      enum:
        - completed
        - in_progress
        - failed
        - scheduled
    PostTrainingJobStatusResponse:
      type: object
      properties:
        job_uuid:
          type: string
        status:
          $ref: '#/components/schemas/JobStatus'
        scheduled_at:
          type: string
          format: date-time
        started_at:
          type: string
          format: date-time
        completed_at:
          type: string
          format: date-time
        resources_allocated:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        checkpoints:
          type: array
          items:
            $ref: '#/components/schemas/Checkpoint'
      additionalProperties: false
      required:
        - job_uuid
        - status
        - checkpoints
      title: Status of a finetuning job.
    ListPostTrainingJobsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            type: object
            properties:
              job_uuid:
                type: string
            additionalProperties: false
            required:
              - job_uuid
      additionalProperties: false
      required:
        - data
    VectorDB:
      type: object
      properties:
        identifier:
          type: string
        provider_resource_id:
          type: string
        provider_id:
          type: string
        type:
          type: string
          const: vector_db
          default: vector_db
        embedding_model:
          type: string
        embedding_dimension:
          type: integer
      additionalProperties: false
      required:
        - identifier
        - provider_resource_id
        - provider_id
        - type
        - embedding_model
        - embedding_dimension
    HealthInfo:
      type: object
      properties:
        status:
          type: string
      additionalProperties: false
      required:
        - status
    RAGDocument:
      type: object
      properties:
        document_id:
          type: string
        content:
          oneOf:
            - type: string
            - $ref: '#/components/schemas/InterleavedContentItem'
            - type: array
              items:
                $ref: '#/components/schemas/InterleavedContentItem'
            - $ref: '#/components/schemas/URL'
        mime_type:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - document_id
        - content
        - metadata
    InsertRequest:
      type: object
      properties:
        documents:
          type: array
          items:
            $ref: '#/components/schemas/RAGDocument'
        vector_db_id:
          type: string
        chunk_size_in_tokens:
          type: integer
      additionalProperties: false
      required:
        - documents
        - vector_db_id
        - chunk_size_in_tokens
    InsertChunksRequest:
      type: object
      properties:
        vector_db_id:
          type: string
        chunks:
          type: array
          items:
            type: object
            properties:
              content:
                $ref: '#/components/schemas/InterleavedContent'
              metadata:
                type: object
                additionalProperties:
                  oneOf:
                    - type: 'null'
                    - type: boolean
                    - type: number
                    - type: string
                    - type: array
                    - type: object
            additionalProperties: false
            required:
              - content
              - metadata
        ttl_seconds:
          type: integer
      additionalProperties: false
      required:
        - vector_db_id
        - chunks
    InvokeToolRequest:
      type: object
      properties:
        tool_name:
          type: string
        kwargs:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - tool_name
        - kwargs
    ToolInvocationResult:
      type: object
      properties:
        content:
          $ref: '#/components/schemas/InterleavedContent'
        error_message:
          type: string
        error_code:
          type: integer
      additionalProperties: false
      required:
        - content
    ListDatasetsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Dataset'
      additionalProperties: false
      required:
        - data
    ListEvalTasksResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/EvalTask'
      additionalProperties: false
      required:
        - data
    ListModelsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Model'
      additionalProperties: false
      required:
        - data
    ProviderInfo:
      type: object
      properties:
        api:
          type: string
        provider_id:
          type: string
        provider_type:
          type: string
      additionalProperties: false
      required:
        - api
        - provider_id
        - provider_type
    ListProvidersResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/ProviderInfo'
      additionalProperties: false
      required:
        - data
    RouteInfo:
      type: object
      properties:
        route:
          type: string
        method:
          type: string
        provider_types:
          type: array
          items:
            type: string
      additionalProperties: false
      required:
        - route
        - method
        - provider_types
    ListRoutesResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/RouteInfo'
      additionalProperties: false
      required:
        - data
    ListScoringFunctionsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/ScoringFn'
      additionalProperties: false
      required:
        - data
    ListShieldsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Shield'
      additionalProperties: false
      required:
        - data
    ListToolGroupsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/ToolGroup'
      additionalProperties: false
      required:
        - data
    ListToolsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Tool'
      additionalProperties: false
      required:
        - data
    ListVectorDBsResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/VectorDB'
      additionalProperties: false
      required:
        - data
    Event:
      oneOf:
        - $ref: '#/components/schemas/UnstructuredLogEvent'
        - $ref: '#/components/schemas/MetricEvent'
        - $ref: '#/components/schemas/StructuredLogEvent'
      discriminator:
        propertyName: type
        mapping:
          unstructured_log: '#/components/schemas/UnstructuredLogEvent'
          metric: '#/components/schemas/MetricEvent'
          structured_log: '#/components/schemas/StructuredLogEvent'
    LogSeverity:
      type: string
      enum:
        - verbose
        - debug
        - info
        - warn
        - error
        - critical
    MetricEvent:
      type: object
      properties:
        trace_id:
          type: string
        span_id:
          type: string
        timestamp:
          type: string
          format: date-time
        attributes:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        type:
          type: string
          const: metric
          default: metric
        metric:
          type: string
        value:
          oneOf:
            - type: integer
            - type: number
        unit:
          type: string
      additionalProperties: false
      required:
        - trace_id
        - span_id
        - timestamp
        - type
        - metric
        - value
        - unit
    SpanEndPayload:
      type: object
      properties:
        type:
          type: string
          const: span_end
          default: span_end
        status:
          $ref: '#/components/schemas/SpanStatus'
      additionalProperties: false
      required:
        - type
        - status
    SpanStartPayload:
      type: object
      properties:
        type:
          type: string
          const: span_start
          default: span_start
        name:
          type: string
        parent_span_id:
          type: string
      additionalProperties: false
      required:
        - type
        - name
    StructuredLogEvent:
      type: object
      properties:
        trace_id:
          type: string
        span_id:
          type: string
        timestamp:
          type: string
          format: date-time
        attributes:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        type:
          type: string
          const: structured_log
          default: structured_log
        payload:
          $ref: '#/components/schemas/StructuredLogPayload'
      additionalProperties: false
      required:
        - trace_id
        - span_id
        - timestamp
        - type
        - payload
    StructuredLogPayload:
      oneOf:
        - $ref: '#/components/schemas/SpanStartPayload'
        - $ref: '#/components/schemas/SpanEndPayload'
      discriminator:
        propertyName: type
        mapping:
          span_start: '#/components/schemas/SpanStartPayload'
          span_end: '#/components/schemas/SpanEndPayload'
    UnstructuredLogEvent:
      type: object
      properties:
        trace_id:
          type: string
        span_id:
          type: string
        timestamp:
          type: string
          format: date-time
        attributes:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        type:
          type: string
          const: unstructured_log
          default: unstructured_log
        message:
          type: string
        severity:
          $ref: '#/components/schemas/LogSeverity'
      additionalProperties: false
      required:
        - trace_id
        - span_id
        - timestamp
        - type
        - message
        - severity
    LogEventRequest:
      type: object
      properties:
        event:
          $ref: '#/components/schemas/Event'
        ttl_seconds:
          type: integer
      additionalProperties: false
      required:
        - event
        - ttl_seconds
    DPOAlignmentConfig:
      type: object
      properties:
        reward_scale:
          type: number
        reward_clip:
          type: number
        epsilon:
          type: number
        gamma:
          type: number
      additionalProperties: false
      required:
        - reward_scale
        - reward_clip
        - epsilon
        - gamma
    DataConfig:
      type: object
      properties:
        dataset_id:
          type: string
        batch_size:
          type: integer
        shuffle:
          type: boolean
        data_format:
          $ref: '#/components/schemas/DatasetFormat'
        validation_dataset_id:
          type: string
        packed:
          type: boolean
          default: false
        train_on_input:
          type: boolean
          default: false
      additionalProperties: false
      required:
        - dataset_id
        - batch_size
        - shuffle
        - data_format
    DatasetFormat:
      type: string
      enum:
        - instruct
        - dialog
    EfficiencyConfig:
      type: object
      properties:
        enable_activation_checkpointing:
          type: boolean
          default: false
        enable_activation_offloading:
          type: boolean
          default: false
        memory_efficient_fsdp_wrap:
          type: boolean
          default: false
        fsdp_cpu_offload:
          type: boolean
          default: false
      additionalProperties: false
    OptimizerConfig:
      type: object
      properties:
        optimizer_type:
          $ref: '#/components/schemas/OptimizerType'
        lr:
          type: number
        weight_decay:
          type: number
        num_warmup_steps:
          type: integer
      additionalProperties: false
      required:
        - optimizer_type
        - lr
        - weight_decay
        - num_warmup_steps
    OptimizerType:
      type: string
      enum:
        - adam
        - adamw
        - sgd
    TrainingConfig:
      type: object
      properties:
        n_epochs:
          type: integer
        max_steps_per_epoch:
          type: integer
        gradient_accumulation_steps:
          type: integer
        max_validation_steps:
          type: integer
        data_config:
          $ref: '#/components/schemas/DataConfig'
        optimizer_config:
          $ref: '#/components/schemas/OptimizerConfig'
        efficiency_config:
          $ref: '#/components/schemas/EfficiencyConfig'
        dtype:
          type: string
          default: bf16
      additionalProperties: false
      required:
        - n_epochs
        - max_steps_per_epoch
        - gradient_accumulation_steps
        - max_validation_steps
        - data_config
        - optimizer_config
    PreferenceOptimizeRequest:
      type: object
      properties:
        job_uuid:
          type: string
        finetuned_model:
          type: string
        algorithm_config:
          $ref: '#/components/schemas/DPOAlignmentConfig'
        training_config:
          $ref: '#/components/schemas/TrainingConfig'
        hyperparam_search_config:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        logger_config:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - job_uuid
        - finetuned_model
        - algorithm_config
        - training_config
        - hyperparam_search_config
        - logger_config
    PostTrainingJob:
      type: object
      properties:
        job_uuid:
          type: string
      additionalProperties: false
      required:
        - job_uuid
    DefaultRAGQueryGeneratorConfig:
      type: object
      properties:
        type:
          type: string
          const: default
          default: default
        separator:
          type: string
          default: ' '
      additionalProperties: false
      required:
        - type
        - separator
    LLMRAGQueryGeneratorConfig:
      type: object
      properties:
        type:
          type: string
          const: llm
          default: llm
        model:
          type: string
        template:
          type: string
      additionalProperties: false
      required:
        - type
        - model
        - template
    RAGQueryConfig:
      type: object
      properties:
        query_generator_config:
          $ref: '#/components/schemas/RAGQueryGeneratorConfig'
        max_tokens_in_context:
          type: integer
          default: 4096
        max_chunks:
          type: integer
          default: 5
      additionalProperties: false
      required:
        - query_generator_config
        - max_tokens_in_context
        - max_chunks
    RAGQueryGeneratorConfig:
      oneOf:
        - $ref: '#/components/schemas/DefaultRAGQueryGeneratorConfig'
        - $ref: '#/components/schemas/LLMRAGQueryGeneratorConfig'
      discriminator:
        propertyName: type
        mapping:
          default: '#/components/schemas/DefaultRAGQueryGeneratorConfig'
          llm: '#/components/schemas/LLMRAGQueryGeneratorConfig'
    QueryRequest:
      type: object
      properties:
        content:
          $ref: '#/components/schemas/InterleavedContent'
        vector_db_ids:
          type: array
          items:
            type: string
        query_config:
          $ref: '#/components/schemas/RAGQueryConfig'
      additionalProperties: false
      required:
        - content
        - vector_db_ids
    RAGQueryResult:
      type: object
      properties:
        content:
          $ref: '#/components/schemas/InterleavedContent'
      additionalProperties: false
    QueryChunksRequest:
      type: object
      properties:
        vector_db_id:
          type: string
        query:
          $ref: '#/components/schemas/InterleavedContent'
        params:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - vector_db_id
        - query
    QueryChunksResponse:
      type: object
      properties:
        chunks:
          type: array
          items:
            type: object
            properties:
              content:
                $ref: '#/components/schemas/InterleavedContent'
              metadata:
                type: object
                additionalProperties:
                  oneOf:
                    - type: 'null'
                    - type: boolean
                    - type: number
                    - type: string
                    - type: array
                    - type: object
            additionalProperties: false
            required:
              - content
              - metadata
        scores:
          type: array
          items:
            type: number
      additionalProperties: false
      required:
        - chunks
        - scores
    QueryCondition:
      type: object
      properties:
        key:
          type: string
        op:
          $ref: '#/components/schemas/QueryConditionOp'
        value:
          oneOf:
            - type: 'null'
            - type: boolean
            - type: number
            - type: string
            - type: array
            - type: object
      additionalProperties: false
      required:
        - key
        - op
        - value
    QueryConditionOp:
      type: string
      enum:
        - eq
        - ne
        - gt
        - lt
    QuerySpansResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Span'
      additionalProperties: false
      required:
        - data
    QueryTracesResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Trace'
      additionalProperties: false
      required:
        - data
    RegisterDatasetRequest:
      type: object
      properties:
        dataset_id:
          type: string
        dataset_schema:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ParamType'
        url:
          $ref: '#/components/schemas/URL'
        provider_dataset_id:
          type: string
        provider_id:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - dataset_id
        - dataset_schema
        - url
    RegisterEvalTaskRequest:
      type: object
      properties:
        eval_task_id:
          type: string
        dataset_id:
          type: string
        scoring_functions:
          type: array
          items:
            type: string
        provider_eval_task_id:
          type: string
        provider_id:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - eval_task_id
        - dataset_id
        - scoring_functions
    RegisterModelRequest:
      type: object
      properties:
        model_id:
          type: string
        provider_model_id:
          type: string
        provider_id:
          type: string
        metadata:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        model_type:
          $ref: '#/components/schemas/ModelType'
      additionalProperties: false
      required:
        - model_id
    RegisterScoringFunctionRequest:
      type: object
      properties:
        scoring_fn_id:
          type: string
        description:
          type: string
        return_type:
          $ref: '#/components/schemas/ParamType'
        provider_scoring_fn_id:
          type: string
        provider_id:
          type: string
        params:
          $ref: '#/components/schemas/ScoringFnParams'
      additionalProperties: false
      required:
        - scoring_fn_id
        - description
        - return_type
    RegisterShieldRequest:
      type: object
      properties:
        shield_id:
          type: string
        provider_shield_id:
          type: string
        provider_id:
          type: string
        params:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - shield_id
    RegisterToolGroupRequest:
      type: object
      properties:
        toolgroup_id:
          type: string
        provider_id:
          type: string
        mcp_endpoint:
          $ref: '#/components/schemas/URL'
        args:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - toolgroup_id
        - provider_id
    RegisterVectorDbRequest:
      type: object
      properties:
        vector_db_id:
          type: string
        embedding_model:
          type: string
        embedding_dimension:
          type: integer
        provider_id:
          type: string
        provider_vector_db_id:
          type: string
      additionalProperties: false
      required:
        - vector_db_id
        - embedding_model
    RunEvalRequest:
      type: object
      properties:
        task_config:
          $ref: '#/components/schemas/EvalTaskConfig'
      additionalProperties: false
      required:
        - task_config
    Job:
      type: object
      properties:
        job_id:
          type: string
      additionalProperties: false
      required:
        - job_id
    RunShieldRequest:
      type: object
      properties:
        shield_id:
          type: string
        messages:
          type: array
          items:
            $ref: '#/components/schemas/Message'
        params:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - shield_id
        - messages
        - params
    RunShieldResponse:
      type: object
      properties:
        violation:
          $ref: '#/components/schemas/SafetyViolation'
      additionalProperties: false
    SaveSpansToDatasetRequest:
      type: object
      properties:
        attribute_filters:
          type: array
          items:
            $ref: '#/components/schemas/QueryCondition'
        attributes_to_save:
          type: array
          items:
            type: string
        dataset_id:
          type: string
        max_depth:
          type: integer
      additionalProperties: false
      required:
        - attribute_filters
        - attributes_to_save
        - dataset_id
    ScoreRequest:
      type: object
      properties:
        input_rows:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
        scoring_functions:
          type: object
          additionalProperties:
            oneOf:
              - $ref: '#/components/schemas/ScoringFnParams'
              - type: 'null'
      additionalProperties: false
      required:
        - input_rows
        - scoring_functions
    ScoreResponse:
      type: object
      properties:
        results:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ScoringResult'
      additionalProperties: false
      required:
        - results
    ScoreBatchRequest:
      type: object
      properties:
        dataset_id:
          type: string
        scoring_functions:
          type: object
          additionalProperties:
            oneOf:
              - $ref: '#/components/schemas/ScoringFnParams'
              - type: 'null'
        save_results_dataset:
          type: boolean
      additionalProperties: false
      required:
        - dataset_id
        - scoring_functions
        - save_results_dataset
    ScoreBatchResponse:
      type: object
      properties:
        dataset_id:
          type: string
        results:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/ScoringResult'
      additionalProperties: false
      required:
        - results
    AlgorithmConfig:
      oneOf:
        - $ref: '#/components/schemas/LoraFinetuningConfig'
        - $ref: '#/components/schemas/QATFinetuningConfig'
      discriminator:
        propertyName: type
        mapping:
          LoRA: '#/components/schemas/LoraFinetuningConfig'
          QAT: '#/components/schemas/QATFinetuningConfig'
    LoraFinetuningConfig:
      type: object
      properties:
        type:
          type: string
          const: LoRA
          default: LoRA
        lora_attn_modules:
          type: array
          items:
            type: string
        apply_lora_to_mlp:
          type: boolean
        apply_lora_to_output:
          type: boolean
        rank:
          type: integer
        alpha:
          type: integer
        use_dora:
          type: boolean
          default: false
        quantize_base:
          type: boolean
          default: false
      additionalProperties: false
      required:
        - type
        - lora_attn_modules
        - apply_lora_to_mlp
        - apply_lora_to_output
        - rank
        - alpha
    QATFinetuningConfig:
      type: object
      properties:
        type:
          type: string
          const: QAT
          default: QAT
        quantizer_name:
          type: string
        group_size:
          type: integer
      additionalProperties: false
      required:
        - type
        - quantizer_name
        - group_size
    SupervisedFineTuneRequest:
      type: object
      properties:
        job_uuid:
          type: string
        training_config:
          $ref: '#/components/schemas/TrainingConfig'
        hyperparam_search_config:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        logger_config:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
        model:
          type: string
        checkpoint_dir:
          type: string
        algorithm_config:
          $ref: '#/components/schemas/AlgorithmConfig'
      additionalProperties: false
      required:
        - job_uuid
        - training_config
        - hyperparam_search_config
        - logger_config
        - model
    SyntheticDataGenerateRequest:
      type: object
      properties:
        dialogs:
          type: array
          items:
            $ref: '#/components/schemas/Message'
        filtering_function:
          type: string
          enum:
            - none
            - random
            - top_k
            - top_p
            - top_k_top_p
            - sigmoid
          title: The type of filtering function.
        model:
          type: string
      additionalProperties: false
      required:
        - dialogs
        - filtering_function
    SyntheticDataGenerationResponse:
      type: object
      properties:
        synthetic_data:
          type: array
          items:
            type: object
            additionalProperties:
              oneOf:
                - type: 'null'
                - type: boolean
                - type: number
                - type: string
                - type: array
                - type: object
        statistics:
          type: object
          additionalProperties:
            oneOf:
              - type: 'null'
              - type: boolean
              - type: number
              - type: string
              - type: array
              - type: object
      additionalProperties: false
      required:
        - synthetic_data
      title: >-
        Response from the synthetic data generation. Batch of (prompt, response, score)
        tuples that pass the threshold.
    VersionInfo:
      type: object
      properties:
        version:
          type: string
      additionalProperties: false
      required:
        - version
  responses: {}
security:
  - Default: []
tags:
  - name: Agents
    description: >-
      Main functionalities provided by this API:

      - Create agents with specific instructions and ability to use tools.

      - Interactions with agents are grouped into sessions ("threads"), and each interaction
      is called a "turn".

      - Agents can be provided with various tools (see the ToolGroups and ToolRuntime
      APIs for more details).

      - Agents can be provided with various shields (see the Safety API for more details).

      - Agents can also use Memory to retrieve information from knowledge bases. See
      the RAG Tool and Vector IO APIs for more details.
    x-displayName: >-
      Agents API for creating and interacting with agentic systems.
  - name: BatchInference (Coming Soon)
  - name: DatasetIO
  - name: Datasets
  - name: Eval
  - name: EvalTasks
  - name: Inference
    description: >-
      This API provides the raw interface to the underlying models. Two kinds of models
      are supported:

      - LLM models: these models generate "raw" and "chat" (conversational) completions.

      - Embedding models: these models generate embeddings to be used for semantic
      search.
    x-displayName: >-
      Llama Stack Inference API for generating completions, chat completions, and
      embeddings.
  - name: Inspect
  - name: Models
  - name: PostTraining (Coming Soon)
  - name: Safety
  - name: Scoring
  - name: ScoringFunctions
  - name: Shields
  - name: SyntheticDataGeneration (Coming Soon)
  - name: Telemetry
  - name: ToolGroups
  - name: ToolRuntime
  - name: VectorDBs
  - name: VectorIO
x-tagGroups:
  - name: Operations
    tags:
      - Agents
      - BatchInference (Coming Soon)
      - DatasetIO
      - Datasets
      - Eval
      - EvalTasks
      - Inference
      - Inspect
      - Models
      - PostTraining (Coming Soon)
      - Safety
      - Scoring
      - ScoringFunctions
      - Shields
      - SyntheticDataGeneration (Coming Soon)
      - Telemetry
      - ToolGroups
      - ToolRuntime
      - VectorDBs
      - VectorIO