GET /doubao/api/v3/contents/generations/tasks/{task_id} for the result.content is a top-level array, not nested under metadata.prompt may be sent at the top level or as a text item inside content.duration as an integer (for example 5 or -1).ratio, resolution, watermark, ...) are top-level.| Scenario | content items |
|---|---|
| Text to video | text |
| Image to video, first frame | text + image_url with role: first_frame |
| Image to video, first and last frame | text + two image_url with role: first_frame and role: last_frame |
| Multimodal reference | text + any mix of image_url, video_url, audio_url |
| Edit video | text that refers to @video 1 + video_url, plus omni_reference_task_type: edit |
| Extend video | text that refers to @video 1 + video_url, plus omni_reference_task_type: extend |
data:image/png;base64,.... The format tag must be lowercase.POST /v1/video/generations when you want the OpenAI-compatible request