Hedra — Generate an avatar (talking-head) video

arazzo: 1.0.1
info:
  title: Hedra — Generate an avatar (talking-head) video
  version: 1.0.0
  summary: Upload an image and an audio track, submit an avatar video generation, and poll for the result.
  x-realizes-capability-ids:
  - BC-3700
  x-capability-derivation:
    method: 'deterministic join: sourceDescriptions -> per-tag OpenAPI -> tag/capability edge. No classification at this step.'
    min_confidence: 0.7
    sources:
    - capability_id: BC-3700
      capability_name: Content Origination & Production Management
      spec: hedra-public-api-openapi.yml
      confidence: 0.7
    model: Turbo EA Capabilities by Vincent Verdet — Turbo EA, https://github.com/vincentmakes/turbo-ea-capabilities, CC BY 4.0
sourceDescriptions:
- name: publicApi
  url: ../openapi/hedra-public-api-openapi.yml
  type: openapi
workflows:
- workflowId: generateAvatarVideo
  summary: Reserve+upload an image and an audio asset, submit an avatar video generation, poll to completion.
  description: >-
    Uses Hedra's Character-3 / Avatar model. Uploads a portrait image and a driving audio track,
    then submits a video generation conditioned on both (start_keyframe_id + audio_id).
  inputs:
    type: object
    required: [api_key, image_name, audio_name, model_slug]
    properties:
      api_key: {type: string}
      image_name: {type: string}
      audio_name: {type: string}
      model_slug: {type: string}
  steps:
  - stepId: reserveImage
    operationId: create_asset_public_assets_post
    parameters:
    - {name: X-API-Key, in: header, value: $inputs.api_key}
    requestBody:
      contentType: application/json
      payload:
        name: $inputs.image_name
        type: image
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      image_asset_id: $response.body#/id
  - stepId: uploadImage
    operationId: upload_asset_public_assets__id__upload_post
    parameters:
    - {name: X-API-Key, in: header, value: $inputs.api_key}
    - {name: id, in: path, value: $steps.reserveImage.outputs.image_asset_id}
    successCriteria:
    - condition: $statusCode == 200
  - stepId: reserveAudio
    operationId: create_asset_public_assets_post
    parameters:
    - {name: X-API-Key, in: header, value: $inputs.api_key}
    requestBody:
      contentType: application/json
      payload:
        name: $inputs.audio_name
        type: audio
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      audio_asset_id: $response.body#/id
  - stepId: uploadAudio
    operationId: upload_asset_public_assets__id__upload_post
    parameters:
    - {name: X-API-Key, in: header, value: $inputs.api_key}
    - {name: id, in: path, value: $steps.reserveAudio.outputs.audio_asset_id}
    successCriteria:
    - condition: $statusCode == 200
  - stepId: submitAvatar
    operationId: generate_asset_public_generations_post
    parameters:
    - {name: X-API-Key, in: header, value: $inputs.api_key}
    requestBody:
      contentType: application/json
      payload:
        type: video
        model_slug: $inputs.model_slug
        start_keyframe_id: $steps.reserveImage.outputs.image_asset_id
        audio_id: $steps.reserveAudio.outputs.audio_asset_id
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      generation_id: $response.body#/id
  - stepId: pollStatus
    operationId: get_status_public_generations__generation_id__status_get
    parameters:
    - {name: X-API-Key, in: header, value: $inputs.api_key}
    - {name: generation_id, in: path, value: $steps.submitAvatar.outputs.generation_id}
    successCriteria:
    - condition: $statusCode == 200
    - context: $response.body
      condition: $.status == "complete"
      type: jsonpath
    outputs:
      url: $response.body#/url
  outputs:
    generation_id: $steps.submitAvatar.outputs.generation_id
    url: $steps.pollStatus.outputs.url