Seedance video generation

Seedance 2.0 submits async video tasks through the official generation endpoint. It supports text-to-video, first-frame / first-and-last-frame, reference image/video/audio, video continuation, video editing, and multimodal composition.

POST /v1/video/generations

Asset library items can be referenced in as . On successful submission, a task_id is returned; poll with Task query.

Available models

Headers

Name	Type	Required	Description
	string	Yes	Request authentication. Use a Bearer Token, for example:
	string	Yes	Request content type. Must be set to

Request body

Name	Type	Required	Description
	string	Yes	Model name
	array<object>	Yes	Multimodal input array; order affects role assignment
	object	No	Video generation parameters; all optional

content items

Name	Type	Required	Description
	string	Yes	Content type: , , , ,
	string	Conditional	Required when ; prompt text
	string	Conditional	Required when ; public image URL or
	string	Conditional	Required when ; public video URL or asset reference
	string	Conditional	Required when ; public audio URL or asset reference
	string	Conditional	Required when ; draft task ID (must be the only content item)
	string	No	Media role: (first frame), (last frame), (reference image), (reference/source video), (reference audio; requires )

metadata

Name	Type	Required	Description
	integer	No	Duration in seconds. Valid range – or (model decides). Default
	string	No	, , , . Default
	string	No	, , , , . Default
	integer	No	Total frame count; mutually exclusive with . Takes precedence over duration when set
	integer	No	Random seed. Same seed + same input yields similar results
	boolean	No	Lock camera (suppress camera motion). Default
	boolean	No	Bottom-right watermark. Default
	boolean	No	Generate/synthesize audio. Must be when using . Default
	boolean	No	Return last-frame image URL for continuation. Default
	boolean	No	Draft mode (faster, slightly lower quality). Default
	string	No	Service tier. Default
	integer	No	Max execution time in seconds. Range –. Default
	string	No	Completion callback URL

Generation mode reference

Mode	content essentials
Text-to-video	+ optional reference image/video/audio
First-frame image-to-video	+
First-and-last-frame video	+ +
Reference-image video	+
Video continuation	+
Video editing	+ +
Multimodal composition	+ mixed reference types
Asset library reference	Use for each URL

Request example

Text-to-video

{
  "model": "doubao-seedance-2-0-260128",
  "content": [
    {
      "type": "text",
      "text": "First-person fruit tea ad: 0-2s pick an apple; 2-4s cut to pouring into a shaker and shaking"
    },
    {
      "type": "image_url",
      "image_url": { "url": "https://example.com/apple.jpg" },
      "role": "reference_image"
    }
  ],
  "metadata": {
    "duration": 8,
    "resolution": "720p",
    "ratio": "16:9"
  }
}

First-frame image-to-video

{
  "model": "doubao-seedance-2-0-260128",
  "content": [
    {
      "type": "text",
      "text": "Animate naturally from the first frame"
    },
    {
      "type": "image_url",
      "image_url": { "url": "https://example.com/first.jpg" },
      "role": "first_frame"
    }
  ],
  "metadata": {
    "duration": 10,
    "resolution": "720p",
    "ratio": "adaptive"
  }
}

Asset library

{
  "model": "doubao-seedance-2-0-260128",
  "content": [
    {
      "type": "text",
      "text": "First-person fruit tea ad"
    },
    {
      "type": "image_url",
      "image_url": { "url": "asset://asset-xxx" },
      "role": "reference_image"
    },
    {
      "type": "video_url",
      "video_url": { "url": "asset://asset-xxx" },
      "role": "reference_video"
    }
  ],
  "metadata": {
    "duration": 8,
    "resolution": "720p",
    "ratio": "16:9"
  }
}

Response example

200

{
  "task_id": "cgt-20260412163502-x8k2m"
}

Response fields

Name	Type	Description
	string	Task ID for task query

Notes

WARNING

content mixing rules (violations may return 400):

• cannot appear together with / ;
• cannot be the only content input; pair it with at least an image or video;
• must be the only element in the content array.

Asset library

Reference asset library resources in the main via . Workflow:

1. Create an asset group and obtain GroupId;
2. Import a public HTTPS asset into that group (async transcoding / moderation) and obtain task_id;
3. Poll until the asset is ready, then reference the returned AssetId in video requests.

Create asset group (POST /v1/seedance/asset/CreateAssetGroup)

Name	Type	Required	Description
	string	Yes	Asset group name, max 64 characters
	string	No	Asset group description, max 300 characters

curl -X POST https://octopusx.ai/v1/seedance/asset/CreateAssetGroup \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
  "Name": "test",
  "Description": "this is description"
}'

Import asset (POST /v1/seedance/asset/CreateAsset)

Name	Type	Required	Description
	string	Yes	Asset group ID from data.Id returned by Create asset group
	string	Yes	Public HTTPS URL of the asset to import
	string	Yes	Asset type: , ,
	string	Yes	Asset name

On success, an AssetId is returned. Prefer in video requests instead of repeating the public URL.

Get asset (POST /v1/seedance/asset/GetAsset)

Name	Type	Required	Description
	string	Yes	Task ID returned when importing the asset

Poll processing progress and details. Once the asset is ready, use it in generation requests.

Related endpoints

• Video APIs overview
• Seedance task query
• Upload file
• Error codes