arazzo: 1.0.1
info:
  title: Anthropic Upload, Verify, and Download a File
  summary: Upload a file, read back its metadata, and download its content when it is downloadable.
  description: >-
    A Files API round-trip. The workflow uploads a file via multipart form data,
    reads back the stored file metadata to confirm it landed, and branches on
    the downloadable flag to download the raw file content. Every step spells
    out its request inline — including the inline anthropic-version header the
    spec defines — so the flow can be read and executed without opening the
    underlying OpenAPI description.

    Adaptation note: the Messages API description in this repository does not
    define a document source that references an uploaded file_id (its
    DocumentSource only supports base64, text, url, and content), so this
    workflow exercises the file lifecycle directly rather than feeding the
    upload into a message.
  version: 1.0.0
sourceDescriptions:
- name: filesApi
  url: ../openapi/anthropic-files-api-openapi.yml
  type: openapi
workflows:
- workflowId: upload-verify-download-file
  summary: Upload a file, confirm its metadata, and download its content.
  description: >-
    Creates a file, retrieves its metadata to confirm storage, and downloads
    the content when the file is marked downloadable.
  inputs:
    type: object
    required:
    - apiKey
    - file
    properties:
      apiKey:
        type: string
        description: A valid Anthropic API key sent in the x-api-key header.
      file:
        type: string
        description: The binary file contents to upload as multipart form data.
  steps:
  - stepId: uploadFile
    description: >-
      Upload a file to Anthropic's servers for later use, sending the binary as
      multipart form data.
    operationId: createFile
    parameters:
    - name: x-api-key
      in: header
      value: $inputs.apiKey
    - name: anthropic-version
      in: header
      value: "2023-06-01"
    requestBody:
      contentType: multipart/form-data
      payload:
        file: $inputs.file
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      fileId: $response.body#/id
      filename: $response.body#/filename
      downloadable: $response.body#/downloadable
  - stepId: verifyMetadata
    description: >-
      Retrieve the stored file metadata to confirm the upload succeeded and
      inspect its mime type, size, and downloadable flag.
    operationId: getFileMetadata
    parameters:
    - name: x-api-key
      in: header
      value: $inputs.apiKey
    - name: anthropic-version
      in: header
      value: "2023-06-01"
    - name: file_id
      in: path
      value: $steps.uploadFile.outputs.fileId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      mimeType: $response.body#/mime_type
      sizeBytes: $response.body#/size_bytes
      downloadable: $response.body#/downloadable
    onSuccess:
    - name: canDownload
      type: goto
      stepId: downloadContent
      criteria:
      - context: $response.body
        condition: $.downloadable == true
        type: jsonpath
    - name: notDownloadable
      type: end
      criteria:
      - context: $response.body
        condition: $.downloadable == false
        type: jsonpath
  - stepId: downloadContent
    description: >-
      Download the raw content of the uploaded file now that it is confirmed
      downloadable.
    operationId: downloadFileContent
    parameters:
    - name: x-api-key
      in: header
      value: $inputs.apiKey
    - name: anthropic-version
      in: header
      value: "2023-06-01"
    - name: file_id
      in: path
      value: $steps.uploadFile.outputs.fileId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      content: $response.body
  outputs:
    fileId: $steps.uploadFile.outputs.fileId
    mimeType: $steps.verifyMetadata.outputs.mimeType
    content: $steps.downloadContent.outputs.content