Skip to main content

Berry Seedance 2.0 Fast

POST 
/v1/video/generations

Submit a Seedance 2.0 Fast video generation task via the new-api video generation task API.

Platform task API (recommended)

  • Submit: POST /v1/video/generations
  • Query: GET /v1/video/generations/{task_id} or GET /v1/tasks/generations/{task_id}
  • List tasks: GET /v1/tasks/generations
  • Cancel: POST /v1/tasks/generations/{task_id}/cancel

Volcengine-compatible paths (optional)

Same paths as Volcengine docs; auth uses your platform API Key. {task_id} always refers to the platform task ID (e.g. task_xxx), not the upstream Volcengine task ID.

  • Submit: POST /api/v3/contents/generations/tasks (body must include model for routing)
  • Query: GET /api/v3/contents/generations/tasks/{task_id}
  • List: GET /api/v3/contents/generations/tasks
  • Cancel: DELETE /api/v3/contents/generations/tasks/{task_id}

After submission, use task_id from the response. Clients must only use platform task_id, not upstream Seedance 2.0 task IDs.

Supports text-to-video, first-frame image-to-video, first/last-frame image-to-video, and multimodal reference video generation with reference images, videos, audio, synced audio, web search tools, and optional last-frame output.

Request body

Seedance 2.0 Fast aligns with Volcengine official parameters; pass them at the top level without wrapping in metadata:

{
  "model": "doubao-seedance-2-0-fast-260128",
  "prompt": "A shiba inu running on a neon street after rain",
  "ratio": "16:9",
  "resolution": "720p",
  "duration": 5,
  "generate_audio": true
}
  • For multimodal input, pass top-level content; if content includes a text entry, prompt is optional
  • Gateway shortcuts image / images / seconds / size map to image input, duration, and resolution
  • metadata remains supported; if the same field exists at top level and in metadata, metadata wins

Request​

Responses​

Task submitted successfully