Seedance Video API

Seedance video generation uses an asynchronous Video API. Submit a task first, then query its status with the returned task ID. This API is separate from Chat Completions.

Before you start

Create a Modelflare API key and use https://modelflare.dev/v1 as the base URL. The public model names are seedance-2.0 and seedance-2.5.

The available duration and resolution combinations can differ by model. Check the Models & Pricing page before sending a request. Resolution values use 480p, 720p, 1080p, or 4k when available.

Text-to-video request

curl https://modelflare.dev/v1/videos \
  -H "Authorization: Bearer YOUR_MODELFLARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.5",
    "prompt": "A white cat runs through a neon-lit street in the rain, cinematic tracking shot",
    "duration": 4,
    "resolution": "720p"
  }'

Image-to-video request

Add image with a publicly reachable HTTPS image URL. The remaining fields are the same as the text-to-video request.

curl https://modelflare.dev/v1/videos \
  -H "Authorization: Bearer YOUR_MODELFLARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.5",
    "prompt": "The subject slowly turns toward the camera while the camera moves forward",
    "image": "https://example.com/reference.png",
    "duration": 4,
    "resolution": "720p"
  }'

Submission response

A successful submission returns a public task ID. id and task_id identify the same task; save either value for later requests.

{
  "id": "task_example_123",
  "task_id": "task_example_123",
  "object": "video",
  "model": "seedance-2.5",
  "status": "queued"
}

Check task status

Poll the task every few seconds. Waiting states may include queued, pending, processing, or in_progress. Stop when the status becomes completed or failed.

curl https://modelflare.dev/v1/videos/task_example_123 \
  -H "Authorization: Bearer YOUR_MODELFLARE_API_KEY"

Example completed response:

{
  "id": "task_example_123",
  "task_id": "task_example_123",
  "object": "video",
  "model": "seedance-2.5",
  "status": "completed",
  "progress": 100,
  "metadata": {
    "url": "https://example.com/generated-video.mp4"
  }
}

Download the result

After completion, download the video through the content endpoint:

curl -L https://modelflare.dev/v1/videos/task_example_123/content \
  -H "Authorization: Bearer YOUR_MODELFLARE_API_KEY" \
  -o seedance.mp4

Main request fields

• model: required; seedance-2.0 or seedance-2.5.
• prompt: required video description.
• duration: required integer in seconds; the current minimum is 4. seconds is also accepted for compatibility.
• resolution: required output resolution supported by the selected model.
• image: optional public HTTPS image URL; adding it switches the request to image-to-video.

HTTP 400 usually indicates invalid parameters, 401 indicates an invalid API key, 429 indicates a temporary limit, and 5xx errors may be retried with backoff.