Overview and Formats
Seedance 2.0 / 2.5 supports official and generic video formats, with automatic asset detection and upload from HTTPS URLs.
Seedance 2.0 / 2.5 uses asynchronous video tasks: create a task, poll by task ID, then read content.video_url, outputs[0], or generic video_url according to the channel response. See Request and Response Parameters for 2.5 task guidance and response fields.
Assets are uploaded automatically. Pass publicly accessible https:// URLs directly in the image, video, or audio input fields of either request format. ModelSell automatically detects and uploads these assets, so you do not need to call /api/assets/upload beforehand.
| Item | Value |
|---|---|
| Official create endpoint | POST /api/v3/contents/generations/tasks |
| Official polling endpoint | GET /api/v3/contents/generations/tasks/{task_id} |
| Generic video create endpoint | POST /v1/video/generations or POST /v1/videos |
| Asset upload endpoint | POST /api/assets/upload |
| Asset detail endpoint | GET /api/assets/{id} |
| Supported models | doubao-seedance-2-5-260628, doubao-seedance-2-0-260128, doubao-seedance-2-0-fast-260128 |
Use the official format when you need exact content[] modality roles. Use the generic video format when your integration already targets /v1/video/generations or /v1/videos.
To reuse images, videos, or audio by asset ID, you can also upload them manually with /api/assets/upload. After the asset becomes Active, pass it as asset://<asset_id> in either request format. See Asset Upload and Management.