AxonFlow Workflows API

openapi: 3.2.0
info:
  title: Axonflow Workflows API
  version: 11.1.0
  contact:
    name: AxonFlow Support
    url: https://getaxonflow.com/support
  license:
    name: Business Source License 1.1
    url: https://github.com/getaxonflow/axonflow/blob/main/LICENSE
  description: 'Operations tagged Workflows across 2 of this provider''s published API definitions: axonflow-orchestrator-api.yaml, axonflow-orchestrator-openapi.yml. Each path carries the servers of the definition it was published in.'
servers:
- url: https://orchestrator.getaxonflow.com
  description: Production (SaaS)
- url: http://localhost:8081
  description: Local Development
tags:
- name: Workflows
  description: Workflow execution engine
paths:
  /api/v1/workflows/execute:
    post:
      tags:
      - Workflows
      summary: Execute a workflow
      description: Execute a defined workflow with input parameters
      operationId: executeWorkflow
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/WorkflowExecuteRequest'
            example:
              workflow:
                metadata:
                  name: data-analysis-workflow
                  description: Analyze sales data
                spec:
                  steps:
                  - name: fetch_data
                    type: mcp_query
                    connector: postgres
                    query: SELECT * FROM sales
                  - name: analyze
                    type: llm
                    prompt: 'Analyze the sales data: {{fetch_data.result}}'
              input:
                start_date: '2025-01-01'
                end_date: '2025-01-15'
              user:
                id: 123
                email: analyst@company.com
                role: analyst
                tenant_id: tenant-abc
      responses:
        '200':
          description: 'Workflow execution result: every step ran, each decided by the

            organization''s policy before it ran (#4382). The returned `id` is

            where a caller first observes the `wfe_` prefix introduced in

            #3442, so an integration that matches run ids on `wf_` breaks

            here rather than on a later GET.


            Since v11.1.0 every step is decided on every deployment, and

            `AXONFLOW_HITL_ENABLED` is ignored. Before it, steps were decided

            only when that variable was `true`, and a run paused by an approval

            answered `200` with `status: "paused"` (or `202` when the approval

            could not be created). Neither pause exists any more: a step that

            requires an approval is withheld (`403`,

            `approval_requires_durable_record`). A plan run in `confirm` or

            `step` mode holds its steps durably.

            '
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/WorkflowExecution'
              example:
                id: wfe_1787399674_6t245pvn
                workflow_name: data-analysis-workflow
                status: completed
                steps:
                - name: fetch_data
                  status: completed
                  process_time: 412ms
                - name: analyze
                  status: completed
                  process_time: 1.83s
                output:
                  summary: Sales rose 12% over the window.
        '400':
          $ref: '#/components/responses/BadRequest'
        '401':
          description: 'The authenticated tenancy scope is missing or unusable. The

            handler resolves the gateway-stamped `X-Org-ID` /

            `X-Tenant-ID` pair before anything else and refuses when

            either dimension is absent, blank, or the unowned-org

            sentinel, with a message opening `Unauthorized: an

            authenticated tenant scope is required` and naming the two

            headers the AxonFlow Agent gateway stamps. The resolved scope

            then overwrites the body''s `user.tenant_id` / `user.org_id`

            and becomes the tenancy stamped on every row this call

            creates. A second stamped-row write guard sits at the row

            boundary (#3066), but it is defense in depth: the scope

            resolution above already guarantees both dimensions, so that

            guard''s distinct refusal message is not reachable on this

            endpoint.

            '
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '403':
          description: 'Refused, with one of two bodies.


            A step was refused before it ran (#4382). Every step is decided

            by the organization''s policy before it runs, on every deployment.

            The body is a `MultiAgentStepRefusal`: `code` is

            `execution_blocked` when a policy blocks the step, or

            `approval_requires_durable_record` when a policy requires an

            approval (this route keeps no durable approval record, so it

            withholds the step; run the plan in `confirm` or `step` mode,

            whose holds are durable), or `route_refused` when the

            organization''s route rows refuse an `llm-call` step''s call (#4249;

            `policy` is the route''s reason). No later step runs, and

            `soft_failure_tolerance` does not absorb a refusal.


            Or the request body carries a cross-tenant claim: a

            `user.tenant_id` naming a tenancy other than the

            authenticated one is refused with `Forbidden: user.tenant_id

            does not name the authenticated tenancy`. Omitting the body

            field is not an error; it is overwritten from the

            authenticated scope.

            '
          content:
            application/json:
              schema:
                oneOf:
                - $ref: '#/components/schemas/MultiAgentStepRefusal'
                - $ref: '#/components/schemas/ErrorResponse'
        '413':
          $ref: '#/components/responses/StepRequestTooLargeFlat'
        '500':
          $ref: '#/components/responses/InternalError'
        '503':
          description: 'The orchestrator''s workflow engine decides no step (its step

            gate is not wired), so nothing runs. A serving orchestrator

            wires it at boot on every deployment.

            '
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
    servers:
    - url: https://orchestrator.getaxonflow.com
      description: Production (SaaS)
    - url: http://localhost:8081
      description: Local Development
  /api/v1/workflows/executions/{id}:
    get:
      tags:
      - Workflows
      summary: Get workflow execution
      description: Get details of a specific workflow execution
      operationId: getWorkflowExecution
      parameters:
      - name: id
        in: path
        required: true
        description: 'In-process declarative workflow-engine execution ID, as returned by POST /api/v1/workflows/execute. Since #3442 these carry the `wfe_` prefix; a governed control-plane `wf_` workflow_id is not resolvable here.'
        schema:
          type: string
        example: wfe_1787399674_6t245pvn
      responses:
        '200':
          description: Workflow execution details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/WorkflowExecution'
        '404':
          description: Execution not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
    servers:
    - url: https://orchestrator.getaxonflow.com
      description: Production (SaaS)
    - url: http://localhost:8081
      description: Local Development
  /api/v1/workflows/executions:
    get:
      tags:
      - Workflows
      summary: List workflow executions
      description: List recent workflow executions
      operationId: listWorkflowExecutions
      parameters:
      - name: limit
        in: query
        description: Maximum number of executions to return
        schema:
          type: integer
          default: 10
          minimum: 1
          maximum: 100
      responses:
        '200':
          description: List of executions
          content:
            application/json:
              schema:
                type: object
                properties:
                  executions:
                    type: array
                    items:
                      $ref: '#/components/schemas/WorkflowExecution'
                  count:
                    type: integer
    servers:
    - url: https://orchestrator.getaxonflow.com
      description: Production (SaaS)
    - url: http://localhost:8081
      description: Local Development
  /api/v1/workflows/executions/tenant/{tenant_id}:
    get:
      tags:
      - Workflows
      summary: Get tenant workflow executions
      description: Get workflow executions for a specific tenant
      operationId: getTenantWorkflowExecutions
      parameters:
      - name: tenant_id
        in: path
        required: true
        description: Tenant identifier
        schema:
          type: string
      responses:
        '200':
          description: Tenant workflow executions
          content:
            application/json:
              schema:
                type: object
                properties:
                  tenant_id:
                    type: string
                  count:
                    type: integer
                  executions:
                    type: array
                    items:
                      $ref: '#/components/schemas/WorkflowExecution'
    servers:
    - url: https://orchestrator.getaxonflow.com
      description: Production (SaaS)
    - url: http://localhost:8081
      description: Local Development
  /api/v1/workflows/executions/{id}/hitl-status:
    get:
      tags:
      - Workflows
      summary: Human-in-the-loop status for a workflow execution
      description: 'Reported whether an execution was paused, in memory, awaiting human

        approval. The multi-agent execute routes pause nothing since v11.1.0

        (#4382), and the in-memory engine this route read is retired (#4249), so

        every id answers 404 `Execution not found`. A multi-agent step is held

        only in confirm or step mode; read its approval state through the

        workflow control plane. The route is removed in v12.0.0.'
      operationId: getHITLExecutionStatus
      parameters:
      - name: id
        in: path
        required: true
        schema:
          type: string
      - name: X-Org-ID
        in: header
        required: true
        schema:
          type: string
      - name: X-Tenant-ID
        in: header
        required: true
        schema:
          type: string
      responses:
        '400':
          $ref: '#/components/responses/BadRequest'
        '404':
          $ref: '#/components/responses/NotFound'
    servers:
    - url: https://orchestrator.getaxonflow.com
      description: Production (SaaS)
    - url: http://localhost:8081
      description: Local Development
components:
  schemas:
    WorkflowExecution:
      type: object
      properties:
        id:
          type: string
          description: 'In-process declarative workflow-engine run identifier. Since #3442 these carry the `wfe_` prefix, not `wf_`: this engine runs a spec handed to it in the request body and appears in `workflows`, `workflow_steps` and `execution_history` nowhere at all, so it is a different thing from a governed control-plane workflow and no longer shares that workflow''s prefix. No component derives meaning from a `wfe_` id; the one place in the platform that parses id prefixes is the unified-executions resolver, which dispatches on `wf_`/`wcp_`/`plan_`, and a `wfe_` id deliberately no longer matches any of them. A caller that stored one and matches on the `wf_` prefix stops matching. Ids minted before v10.0.0 keep their old prefix; no row is rewritten.'
          example: wfe_1787399674_6t245pvn
        workflow_name:
          type: string
        status:
          type: string
          enum:
          - pending
          - running
          - completed
          - failed
        input:
          type: object
          additionalProperties: true
        start_time:
          type: string
          format: date-time
          description: Earlier revisions of this spec named this field `started_at`; the wire has always been `start_time`. Same for `end_time` below, previously misdocumented as `completed_at`.
        end_time:
          type: string
          format: date-time
          description: Omitted while the run is still executing.
        user_context:
          $ref: '#/components/schemas/UserContext'
        error:
          type: string
          description: Present only when the run failed.
        steps:
          type: array
          items:
            type: object
            properties:
              name:
                type: string
              status:
                type: string
                enum:
                - pending
                - running
                - completed
                - failed
                - skipped
              input:
                type: object
                additionalProperties: true
              output:
                type: object
                additionalProperties: true
              start_time:
                type: string
                format: date-time
              end_time:
                type: string
                format: date-time
                description: Omitted while the step is still executing.
              error:
                type: string
                description: Present only when the step failed.
              process_time:
                type: string
        output:
          type: object
          additionalProperties: true
    Workflow:
      type: object
      required:
      - metadata
      - spec
      properties:
        metadata:
          type: object
          required:
          - name
          properties:
            name:
              type: string
            description:
              type: string
        spec:
          type: object
          required:
          - steps
          properties:
            steps:
              type: array
              items:
                $ref: '#/components/schemas/WorkflowStep'
    MultiAgentStepRefusal:
      description: 'A multi-agent step refused before it ran (#4382), on

        `POST /api/v1/workflows/execute` and `POST /api/v1/plan/execute`.

        This is the flat `ErrorResponse` envelope with a machine-readable

        `code` and the refusing policy added. It is a refinement of that

        shape, not a fourth one, as `LLMProviderAPIError` refines the coded

        shape.

        '
      allOf:
      - $ref: '#/components/schemas/ErrorResponse'
      - type: object
        properties:
          error:
            type: string
            example: 'execution blocked by policy: ceiling.no_tool_calls'
          code:
            type: string
            enum:
            - execution_blocked
            - approval_requires_durable_record
            - route_refused
            description: '`execution_blocked`: a policy blocks the step. `approval_requires_durable_record`: a policy requires an approval, and this execution path keeps no durable approval record, so the step is withheld; `confirm` and `step` mode hold durably. `route_refused`: the organization''s route rows refuse an `llm-call` step''s call before any provider is called, and `policy` is the route''s reason (`no_compliant_provider`, `segment_not_established`, `segment_resolution_failed` or `decision_enforcement_unavailable`).'
          policy:
            type: string
            description: The policy, or the cause, that refused the step.
          reason:
            type: string
            description: Why the step was refused.
        required:
        - code
        - policy
        - reason
    WorkflowStep:
      type: object
      required:
      - name
      - type
      properties:
        name:
          type: string
        type:
          type: string
          enum:
          - mcp_query
          - llm
          - api_call
          - transform
        connector:
          type: string
        query:
          type: string
        prompt:
          type: string
        dependencies:
          type: array
          items:
            type: string
    UserContext:
      type: object
      properties:
        id:
          type: integer
        email:
          type: string
        role:
          type: string
        region:
          type: string
          description: User's region, read by geo-based routing policies.
        permissions:
          type: array
          items:
            type: string
        tenant_id:
          type: string
        org_id:
          type: string
          description: Organisation for multi-tenant isolation, populated from the X-Org-ID header the agent stamps on the trusted hop.
    WorkflowExecuteRequest:
      type: object
      required:
      - workflow
      properties:
        workflow:
          $ref: '#/components/schemas/Workflow'
        input:
          type: object
          additionalProperties: true
        user:
          $ref: '#/components/schemas/UserContext'
    ErrorResponse:
      type: object
      description: 'The FLAT error envelope: `{success, error}`. This is what

        `sendErrorResponse` emits, which is the orchestrator''s dominant error

        writer (240 call sites), so it is the shape of every error from the

        core request, audit, plan, workflow, execution and connector surfaces.


        It is one of THREE error SHAPES this document describes. See

        `CodedErrorResponse` and `TripletErrorResponse` for the other two, and

        the note on `components.responses` for why there is more than one.

        `LLMProviderAPIError` is a code-constrained refinement of the coded

        shape, not a fourth shape.


        This paragraph said "TWO" until issue #3941. `TripletErrorResponse` was

        added by the #3901 reconciliation and this sentence was not updated with

        it, so the document undercounted its own families — which is the same

        defect one level up as the operations that named the wrong one.

        '
      properties:
        success:
          type: boolean
          example: false
        error:
          type: string
          description: Human-readable message. There is no machine-readable code on this envelope.
      required:
      - success
      - error
  responses:
    BadRequest:
      description: Invalid request
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/ErrorResponse'
          example:
            success: false
            error: Invalid request body
    InternalError:
      description: Internal server error
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/ErrorResponse'
          example:
            success: false
            error: Internal server error
    StepRequestTooLargeFlat:
      description: 'The request body is over 1 MiB (1,048,576 bytes), refused whole before

        it is decoded, in this route''s `{success, error}` envelope: `error`

        begins `request_too_large: `. Nothing is gated, planned or run (#4249).

        '
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/ErrorResponse'
          example:
            success: false
            error: 'request_too_large: request body exceeds 1048576 bytes; nothing was gated or run'
    NotFound:
      description: Resource not found
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/ErrorResponse'
          example:
            success: false
            error: Resource not found
  securitySchemes:
    basicAuth:
      type: http
      scheme: basic
      description: OAuth2-style client credentials (clientId:clientSecret)
    BearerAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
      description: Enterprise JWT token (see /scripts/generate-jwt.sh)
x-refined-from:
- axonflow-orchestrator-api.yaml
- axonflow-orchestrator-openapi.yml