arazzo: 1.0.1
info:
  title: fal Streaming Inference
  summary: Run a model synchronously over the streaming endpoint to receive progressive output.
  description: >-
    For chatty multimodal models and live previews, fal exposes a synchronous
    streaming endpoint that emits Server-Sent Events as the model produces
    progressive output. This workflow posts a model-specific input to the
    stream endpoint and confirms the SSE stream was opened. Because the
    response is a text/event-stream rather than a JSON document, the workflow
    treats a 200 response as the successful opening of the stream and leaves
    incremental event consumption to the client. Every step spells out its
    request inline so the flow can be read and executed without opening the
    underlying OpenAPI description.
  version: 1.0.0
sourceDescriptions:
- name: falModelApis
  url: ../openapi/fal-model-apis-openapi.yml
  type: openapi
workflows:
- workflowId: stream-inference
  summary: Open a synchronous SSE stream of progressive model output.
  description: >-
    Posts a model-specific input to the streaming endpoint and confirms the
    Server-Sent Events stream was opened so the client can consume progress and
    output events.
  inputs:
    type: object
    required:
    - modelOwner
    - modelName
    - input
    properties:
      modelOwner:
        type: string
        description: Owning organization of the model (e.g. "fal-ai").
      modelName:
        type: string
        description: Model identifier (e.g. "any-llm").
      input:
        type: object
        description: Model-specific JSON input (prompt, image_url, seed, etc.).
  steps:
  - stepId: openStream
    description: >-
      Post the model-specific input to the streaming endpoint and confirm the
      SSE stream was opened with a 200 response. Incremental progress and output
      events are consumed by the client from the open stream.
    operationId: streamRequest
    parameters:
    - name: model_owner
      in: path
      value: $inputs.modelOwner
    - name: model_name
      in: path
      value: $inputs.modelName
    requestBody:
      contentType: application/json
      payload: $inputs.input
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      statusCode: $statusCode
  outputs:
    statusCode: $steps.openStream.outputs.statusCode