> ## Documentation Index
> Fetch the complete documentation index at: https://docs.prisme.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Stream an agent reply via SSE (A2A)

> A2A-compliant streaming send. The request body uses the same
`MessageSendParams` shape as `messages/send`. The response is
delivered as Server-Sent Events (`text/event-stream`): each
event is one of `task.status`, `task.message`, `task.artifact`,
`task.completed`, `task.failed`, `task.canceled`. Clients must
read until they see a terminal event.

The `x-draft-mode: true` header bypasses `published_config`
resolution. The same HITL approval flow as `messages/send`
applies - a streaming task may park in `input-required` until
resolved.

This automation is also bound to the workspace event
`agents.messages.stream` for native event-bus fan-out
(WebSocket consumers); that path is out of scope for OpenAPI.




## OpenAPI

````yaml /api-reference/agent-factory/swagger.yml post /v1/agents/{agentId}/messages/stream
openapi: 3.0.3
info:
  version: 1.0.0
  title: Agent Factory API
  description: |
    Public REST API for the Agent Factory workspace - agents, conversations,
    messages, tools, artifacts, sharing, ratings, and discovery.
    Powered by Prisme.ai's runtime; all endpoints are exposed under each
    workspace's webhook namespace.
  contact:
    name: Prisme.ai
    url: https://prisme.ai
servers:
  - url: https://{host}/v2/workspaces/slug:agent-factory/webhooks
    description: Prisme.ai workspace webhooks
    variables:
      host:
        default: api.studio.prisme.ai
        description: API host (override for self-hosted or sandbox)
security:
  - BearerAuth: []
  - OrgApiKeyAuth: []
tags:
  - name: Agents
    description: Agent CRUD, discovery, AGENTS.md import/export.
  - name: Access
    description: Agent access bindings, sharing, and access requests.
  - name: ApiKeys
    description: Agent-scoped API key management (mint, revoke, rotate).
  - name: Publishing
    description: Publish or discard draft changes on an agent.
  - name: Ratings
    description: User ratings on published agents.
  - name: Profiles
    description: >-
      Agent profiles/presets catalog (simple, workflow, agent_light, agent_full,
      orchestrator).
  - name: Activity
    description: Activity feed for agents (events, errors, lifecycle changes).
  - name: Analytics
    description: Agent usage analytics (series + summary).
  - name: Conversations
    description: Conversations on an agent (CRUD, archive, star).
  - name: Messages
    description: Send messages to an agent (synchronous send + SSE stream).
  - name: Tasks
    description: Async task lifecycle (list, fetch, cancel, resolve, subscribe).
  - name: Artifacts
    description: Generated artifacts (files, code, content) attached to a task.
  - name: Shares
    description: Conversation, message, and artifact share-link snapshots.
  - name: A2A
    description: Agent-to-agent JSON-RPC 2.0 gateway (well-known agent.json + RPC).
  - name: Tools
    description: Per-agent tool catalogue (system tools, MCP servers, function tools).
  - name: Retention
    description: Per-agent and org-wide conversation retention policies.
  - name: Evaluations
    description: Agent evaluation runs and results.
paths:
  /v1/agents/{agentId}/messages/stream:
    post:
      tags:
        - Messages
      summary: Stream an agent reply via SSE (A2A)
      description: |
        A2A-compliant streaming send. The request body uses the same
        `MessageSendParams` shape as `messages/send`. The response is
        delivered as Server-Sent Events (`text/event-stream`): each
        event is one of `task.status`, `task.message`, `task.artifact`,
        `task.completed`, `task.failed`, `task.canceled`. Clients must
        read until they see a terminal event.

        The `x-draft-mode: true` header bypasses `published_config`
        resolution. The same HITL approval flow as `messages/send`
        applies - a streaming task may park in `input-required` until
        resolved.

        This automation is also bound to the workspace event
        `agents.messages.stream` for native event-bus fan-out
        (WebSocket consumers); that path is out of scope for OpenAPI.
      operationId: streamMessage
      parameters:
        - name: agentId
          in: path
          required: true
          schema:
            type: string
            maxLength: 64
        - name: x-draft-mode
          in: header
          required: false
          schema:
            type: boolean
          description: |
            When true, the stream exercises the agent's live (draft)
            config instead of `published_config`. Owners/writers only.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/MessageSendInput'
      responses:
        '200':
          description: |
            Server-Sent Events stream of A2A task lifecycle events.
            Terminates after a `task.completed`, `task.failed`, or
            `task.canceled` event.
          content:
            text/event-stream:
              schema:
                type: string
                description: |
                  Stream of `data: <json>` lines, each carrying an A2A
                  task lifecycle event. See description for event types.
        '400':
          description: Validation error.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '401':
          description: Authentication required.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '403':
          description: Forbidden.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '404':
          description: Agent not found.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '429':
          description: Rate limit exceeded.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
components:
  schemas:
    MessageSendInput:
      type: object
      required:
        - message
      properties:
        message:
          $ref: '#/components/schemas/Message'
    Error:
      type: object
      required:
        - error
        - message
      properties:
        error:
          type: string
          description: >-
            Stable PascalCase or SNAKE_CASE error code (e.g. AgentNotFound,
            VALIDATION_ERROR, FORBIDDEN, METHOD_NOT_ALLOWED).
        message:
          type: string
          description: Human-readable error message.
        details:
          type: object
          description: Optional structured context for the error.
          additionalProperties: true
    Message:
      type: object
      description: A2A v0.3 flat message envelope.
      properties:
        message_id:
          type: string
          maxLength: 256
        role:
          type: string
          enum:
            - user
            - agent
        contextId:
          type: string
          maxLength: 64
        taskId:
          type: string
        parts:
          type: array
          items:
            $ref: '#/components/schemas/MessagePart'
    MessagePart:
      type: object
      description: |
        A2A message part. Discriminated by field presence:
        `TextPart` (`text`), `FilePart` (`url`/`raw`, `mediaType`,
        `filename`), `DataPart` (`data`, `mediaType`).
      additionalProperties: true
  securitySchemes:
    BearerAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
      description: |
        User-bound credential carrying an identity: either a session JWT
        or a user access token (`at:*`) generated from the user settings UI.
        Send as `Authorization: Bearer <token>`.
        Org API keys (`iak_*`) are **not** accepted here - they carry
        no user identity. Use the `x-prismeai-api-key` header instead
        (see `OrgApiKeyAuth`).
    OrgApiKeyAuth:
      type: apiKey
      in: header
      name: x-prismeai-api-key
      description: |
        Organization API key (`iak_{orgSlug}_{uuid}`). Unlike
        `Authorization: Bearer`, this credential is **not** tied to a user
        identity - it is bound to the org and its effective access is
        defined by the scopes / permission rules attached to it (it can
        be restricted to a single project, or kept broader).
        For Agent Factory, these keys can be generated directly from
        the Agent Factory UI (in addition to the AI Governance settings).
        Send as `x-prismeai-api-key: iak_...`.

````