arazzo: 1.0.1
info:
  title: OpenAI Transcribe then Summarize
  summary: Transcribe an audio file, then summarize the transcript with a chat completion.
  description: >-
    Transcribes an uploaded audio file into text, then passes that transcript to
    the chat completions endpoint to produce a concise summary. 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: openaiApi
  url: ../openapi/openai-openapi-master.yml
  type: openapi
workflows:
- workflowId: transcribe-then-summarize
  summary: Transcribe audio and summarize the resulting transcript.
  description: >-
    Sends an audio file to the transcription endpoint, then asks a chat model to
    summarize the returned transcript text.
  inputs:
    type: object
    required:
    - apiKey
    - audioFile
    - transcribeModel
    - chatModel
    properties:
      apiKey:
        type: string
        description: OpenAI API key used as a Bearer token.
      audioFile:
        type: string
        description: The audio file contents to transcribe (binary/multipart field).
      transcribeModel:
        type: string
        description: The transcription model id (e.g. gpt-4o-transcribe or whisper-1).
      chatModel:
        type: string
        description: The chat model id used to summarize the transcript (e.g. gpt-4o-mini).
  steps:
  - stepId: transcribe
    description: Transcribe the uploaded audio file into text.
    operationId: createTranscription
    parameters:
    - name: Authorization
      in: header
      value: "Bearer $inputs.apiKey"
    requestBody:
      contentType: multipart/form-data
      payload:
        file: $inputs.audioFile
        model: $inputs.transcribeModel
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      transcript: $response.body#/text
  - stepId: summarize
    description: Summarize the transcript text with a chat completion.
    operationId: createChatCompletion
    parameters:
    - name: Authorization
      in: header
      value: "Bearer $inputs.apiKey"
    requestBody:
      contentType: application/json
      payload:
        model: $inputs.chatModel
        messages:
        - role: system
          content: You summarize transcripts into a concise paragraph.
        - role: user
          content: $steps.transcribe.outputs.transcript
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      summary: $response.body#/choices/0/message/content
  outputs:
    transcript: $steps.transcribe.outputs.transcript
    summary: $steps.summarize.outputs.summary