Seedance Video

Seedance Video uses the official-compatible video task API. Create a task first, then poll the task endpoint for status and the final video URL.

Endpoints

Capability	Method	Path
Unified create video task	`POST`	`/v1/video/generations`
Unified retrieve video task	`GET`	`/v1/video/generations/{task_id}`
Official-compatible create video task	`POST`	`/api/v3/contents/generations/tasks`
Official-compatible retrieve video task	`GET`	`/api/v3/contents/generations/tasks/{task_id}`
Moderate images	`POST`	`/v1/images/moderations`

Supported Models

Model	Notes
`doubao-seedance-2.0`	Seedance 2.0 standard model

The actual model availability depends on account permissions and platform configuration.

Unified Video API

Use the unified video API when you want the same request flow as other video models. Create a task first, then poll the unified retrieve endpoint for status and the final video URL.

http

POST /v1/video/generations
GET /v1/video/generations/{task_id}

Parameters

Parameter	Required	Description
`model`	Yes	Seedance model name. Use `doubao-seedance-2.0`
`prompt`	Yes	Video generation prompt
`seconds`	No	Duration as a string, such as `"5"`; `duration` is also accepted
`images`	No	Image URL array for reference images or first-frame input
`image`	No	Single image URL, equivalent to a one-item `images` array
`input_reference`	No	Single reference media URL, equivalent to single-image input
`metadata`	No	Seedance-specific options passed through to the task request

Common metadata fields:

Field	Description
`resolution`	Output resolution, such as `720p` or `1080p`; pass it only when the selected model supports it
`ratio`	Aspect ratio, such as `16:9`, `9:16`, or `1:1`
`duration`	Video duration in seconds
`seed`	Random seed
`camera_fixed`	Whether to keep the camera fixed
`watermark`	Whether to add a watermark
`generate_audio`	Whether to generate audio
`content`	Official-compatible content array. Use it when you need roles such as `first_frame`, `last_frame`, or `reference_image`

Text-to-Video Example

bash

curl -X POST https://cubicspaces.cloud/v1/video/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "doubao-seedance-2.0",
    "prompt": "A cinematic aerial shot of a futuristic cubic city at sunrise",
    "seconds": "5",
    "metadata": {
      "resolution": "1080p",
      "ratio": "16:9",
      "watermark": false,
      "generate_audio": true
    }
  }'

Reference Image / First Frame Example

bash

curl -X POST https://cubicspaces.cloud/v1/video/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "doubao-seedance-2.0",
    "prompt": "Use image 1 as the subject reference. Make the product rotate slowly on a clean studio background while preserving its appearance.",
    "seconds": "5",
    "metadata": {
      "resolution": "720p",
      "ratio": "1:1",
      "camera_fixed": true,
      "watermark": false,
      "content": [
        {
          "type": "image_url",
          "image_url": {
            "url": "asset://reviewed-image-asset-id"
          },
          "role": "reference_image"
        }
      ]
    }
  }'

A successful unified create request returns a standard video task object:

json

{
  "id": "video_xxx",
  "task_id": "video_xxx",
  "object": "video",
  "model": "doubao-seedance-2.0",
  "status": "queued",
  "progress": 0,
  "created_at": 1770000000
}

Retrieve the task:

bash

curl https://cubicspaces.cloud/v1/video/generations/video_xxx \
  -H "Authorization: Bearer YOUR_API_KEY"

When the task is completed, read the video URL from metadata.url.

Official-Compatible API

http

POST /api/v3/contents/generations/tasks
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Parameters

Parameter	Required	Description
`model`	Yes	Seedance model name
`content`	Yes	Multimodal content array. It usually contains one text item and optional image, video, or audio assets
`content[].type`	Yes	Content type, such as `text`, `image_url`, `video_url`, or `audio_url`
`content[].text`	No	Text prompt, used when `type` is `text`
`content[].image_url.url`	No	Image URL. It can be a public image URL or an approved `asset://<asset ID>`
`content[].video_url.url`	No	Video URL. It can be a public video URL or an approved `asset://<asset ID>`
`content[].role`	No	Asset role, such as `reference_image`, `first_frame`, or `last_frame`
`duration`	No	Video duration in seconds
`ratio`	No	Aspect ratio, such as `16:9`, `9:16`, or `1:1`
`seed`	No	Random seed
`camera_fixed`	No	Whether to keep the camera fixed
`watermark`	No	Whether to add a watermark
`generate_audio`	No	Whether to generate audio

Request Example

bash

curl -X POST https://cubicspaces.cloud/api/v3/contents/generations/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "doubao-seedance-2.0",
    "content": [
      {
        "type": "text",
        "text": "Use image 1 to generate a 5-second product showcase video."
      },
      {
        "type": "image_url",
        "image_url": {
          "url": "asset://reviewed-image-asset-id"
        },
        "role": "reference_image",
        "duration": 5
      }
    ],
    "ratio": "16:9",
    "duration": 5,
    "generate_audio": true
  }'

A successful create request returns an official-compatible task object. id is the task ID. The response may keep additional task-service fields such as service_tier or execution_expires_after:

json

{
  "id": "task_xxx",
  "model": "doubao-seedance-2.0",
  "status": "running",
  "created_at": 1770000000,
  "updated_at": 1770000000,
  "service_tier": "default",
  "execution_expires_after": 172800
}

Moderate Images

When Seedance uses real-person or library-managed image assets, submit public image URLs to the image moderation endpoint first. After approval and ingestion, use the returned items[].asset_url as content[].image_url.url.

http

POST /v1/images/moderations
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

/v1/assets/moderations uses the same moderation handler.

Request Parameters

Parameter	Required	Description
`model`	Yes	Used to route the request to a Seedance channel. Use `doubao-seedance-2.0`
`images`	No	Image URL array
`image_urls`	No	Image URL array, equivalent to `images`
`image_url`	No	Single image URL
`asset_type`	No	Asset type, defaults to `Image`; keep the default for image moderation

At least one of images, image_urls, or image_url is required. Images must be public http or https URLs. base64 and inline binary content are not supported.

Moderation Limits

Submit at most 50 URLs per request. One image per request is recommended because multiple assets may take longer to process.
Use one asset type per batch. For image moderation, use asset_type: "Image".
Image URLs need supported file extensions: .jpeg, .jpg, .png, .webp, .bmp, .tiff, .tif, .gif, .heic, .heif.
Each image should have an aspect ratio between 0.4 and 2.5, width and height between 300 and 6000 px, and file size below 30 MB.
This endpoint is a synchronous wrapper. The platform waits for moderation and returns a normalized result; if processing is delayed, it polls briefly before returning a timeout error.

Request Example

bash

curl -X POST https://cubicspaces.cloud/v1/images/moderations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "doubao-seedance-2.0",
    "images": [
      "https://example.com/product.png"
    ]
  }'

Response Example

json

{
  "code": "success",
  "message": "",
  "data": {
    "object": "asset_moderation",
    "status": "approved",
    "review_batch_id": "review-batch-id",
    "task_id": "moderation-task-id",
    "items": [
      {
        "source_url": "https://example.com/product.png",
        "asset_url": "asset://reviewed-image-asset-id",
        "submit_review_status": 1,
        "passed": true
      }
    ]
  }
}

When passed is true and submit_review_status is 1, the image is approved. asset_url is usually in the asset://<asset ID> form. Put it in content, and set a role such as reference_image, first_frame, or last_frame.

In prompts, refer to assets with the “asset type + index” format, such as “image 1” or “video 1”. The index starts from 1 within assets of the same type in the content array. Do not refer to assets by Asset ID in the prompt.

Asset codes from the legacy asset system cannot be used directly with Seedance 2.0. Use the asset_url returned by the current moderation endpoint, or build asset://<asset ID> from the new approved asset ID.

Text-to-Video Example

bash

curl -X POST https://cubicspaces.cloud/api/v3/contents/generations/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "doubao-seedance-2.0",
    "content": [
      {
        "type": "text",
        "text": "A cinematic aerial shot of a futuristic cubic city at sunrise"
      }
    ],
    "duration": 5,
    "ratio": "16:9",
    "watermark": false
  }'

Moderated Image-to-Video Example

bash

curl -X POST https://cubicspaces.cloud/api/v3/contents/generations/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "doubao-seedance-2.0",
    "content": [
      {
        "type": "text",
        "text": "Use image 1 as the subject reference. Make the product rotate slowly on a clean studio background while preserving its appearance."
      },
      {
        "type": "image_url",
        "image_url": {
          "url": "asset://reviewed-image-asset-id"
        },
        "role": "reference_image"
      }
    ],
    "duration": 5,
    "ratio": "1:1",
    "camera_fixed": true,
    "watermark": false
  }'

First And Last Frame Example

First and last frames use the first_frame and last_frame roles. Each asset can be either a public URL or an approved asset://<asset ID>.

json

{
  "model": "doubao-seedance-2.0",
  "content": [
    {
      "type": "text",
      "text": "Generate a smooth transition video from image 1 to image 2."
    },
    {
      "type": "image_url",
      "image_url": {
        "url": "asset://first-frame-asset-id"
      },
      "role": "first_frame"
    },
    {
      "type": "image_url",
      "image_url": {
        "url": "asset://last-frame-asset-id"
      },
      "role": "last_frame"
    }
  ],
  "duration": 8,
  "ratio": "16:9",
  "generate_audio": true
}

Video Reference Input

For Seedance 2.0 video reference input, pass a video_url item in content. Real-person or authorized assets can also use asset://<asset ID>.

json

{
  "model": "doubao-seedance-2.0",
  "content": [
    {
      "type": "text",
      "text": "Keep the character style and generate a new action sequence"
    },
    {
      "type": "video_url",
      "video_url": {
        "url": "https://example.com/reference.mp4"
      }
    }
  ],
  "duration": 5,
  "ratio": "16:9"
}

Response And Polling

Save id from the create response and use it to poll the task:

http

GET /api/v3/contents/generations/tasks/{task_id}
Authorization: Bearer YOUR_API_KEY

While the task is running, the response usually looks like:

json

{
  "id": "task_xxx",
  "model": "doubao-seedance-2.0",
  "status": "running",
  "created_at": 1770000000,
  "updated_at": 1770000030
}

When the task completes, the response returns status: "succeeded" and the final video URL in content.video_url:

json

{
  "id": "task_xxx",
  "model": "doubao-seedance-2.0",
  "status": "succeeded",
  "content": {
    "video_url": "https://example.com/generated-video.mp4"
  },
  "created_at": 1770000000,
  "updated_at": 1770000600
}

If the task fails, the response returns status: "failed" and the failure reason in error.message:

json

{
  "id": "task_xxx",
  "model": "doubao-seedance-2.0",
  "status": "failed",
  "error": {
    "code": "task_failed",
    "message": "The request failed because the output video may be related to policy restrictions."
  },
  "created_at": 1770000000,
  "updated_at": 1770000300
}

Dashboard Log Labels

Seedance video generation is an asynchronous task. In dashboard usage logs, refund-related records are split into two labels:

Log label	Meaning
Failure Refund	The task failed and the quota held for that task was returned
Pre-deduction Return	The task succeeded, then the platform settled the final task result and returned the unused held portion

Pre-deduction Return does not mean the video task failed. Use the task polling response fields status, error.message, and content.video_url to determine the actual generation result.

Status Values

Status	Meaning
`queued`	Queued
`running`	Generating
`succeeded`	Completed
`failed`	Failed; inspect `error.message`

Notes

base64 images and inline binary assets are not supported. Upload reference or moderation images to public storage first, then pass the URL.
For moderated/ingested assets, use image_url / video_url / audio_url in content; the URL can be asset://<asset ID>.
In prompts, refer to assets as “image 1 / video 1 / audio 1”; do not use Asset IDs directly.
This documentation lists doubao-seedance-2.0; actual availability follows account permissions and platform configuration.
resolution is only supported by some models. Omit it unless the selected model supports it.

Seedance Video ​

Endpoints ​

Supported Models ​

Unified Video API ​

Parameters ​

Text-to-Video Example ​

Reference Image / First Frame Example ​

Official-Compatible API ​

Parameters ​

Request Example ​

Moderate Images ​

Request Parameters ​

Moderation Limits ​

Request Example ​

Response Example ​

Text-to-Video Example ​

Moderated Image-to-Video Example ​

First And Last Frame Example ​

Video Reference Input ​

Response And Polling ​

Dashboard Log Labels ​

Status Values ​

Notes ​

Seedance Video

Endpoints

Supported Models

Unified Video API

Parameters

Text-to-Video Example

Reference Image / First Frame Example

Official-Compatible API

Parameters

Request Example

Moderate Images

Request Parameters

Moderation Limits

Request Example

Response Example

Text-to-Video Example

Moderated Image-to-Video Example

First And Last Frame Example

Video Reference Input

Response And Polling

Dashboard Log Labels

Status Values

Notes