Seedance 2.0 Fast — text, image, and portrait to video API

Use seedance-2.0-fast when you want Seedance-style motion generation through the same Quotaflow /openai/v1/videos task API. The public call stays unified: create a task, store the returned id, and poll until the task completes.

Seedance 2.0 and Veo are different video model families. seedance-2.0-fast is never used as a fallback for veo-* requests.

Model contract

Call /models with your production key before launch. Video access is key-enabled, so a key without Seedance video access should not show seedance-2.0-fast.

4K Seedance requests should still use the public seedance-2.0-fast id. Quotaflow maps those jobs to the appropriate 4K-capable Seedance 2.0 capacity internally; there is no separate public seedance-2.0-4k model id.

Supported public inputs

• prompt for text-to-video direction.
• image_url, image_urls, images, or reference_images for first-frame, portrait, product, character, or style references.
• audio_url or audio_urls for audio reference workflows when enabled for your key and model.
• duration, durationSeconds, or seconds for duration.
• resolution and aspectRatio / ratio for output shape.
• generate_audio, watermark, and return_last_frame when enabled for your package.

Quotaflow prepares reference media internally when a Seedance workflow needs asset-style inputs. Your application should keep using the public /videos endpoint.

Portrait / identity reference example

curl https://api.quotaflow.ai/openai/v1/videos \
  -H "Authorization: Bearer $QUOTAFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.0-fast",
    "prompt": "Animate this portrait into a natural 4 second speaking-style social clip. Keep the face identity, lighting, and framing consistent.",
    "reference_images": [{ "url": "https://example.com/reference-face.jpg" }],
    "durationSeconds": 4,
    "resolution": "4k",
    "aspectRatio": "9:16",
    "generate_audio": true,
    "watermark": false,
    "return_last_frame": true
  }'

Product reference example

curl https://api.quotaflow.ai/openai/v1/videos \
  -H "Authorization: Bearer $QUOTAFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.0-fast",
    "prompt": "Create a smooth lifestyle product clip from the reference image. Keep the product shape and color consistent, soft studio lighting, slow camera push-in.",
    "image_urls": ["https://example.com/product-reference.png"],
    "durationSeconds": 4,
    "resolution": "720p",
    "aspectRatio": "16:9"
  }'

Poll the task

curl https://api.quotaflow.ai/openai/v1/videos/$TASK_ID \
  -H "Authorization: Bearer $QUOTAFLOW_API_KEY"

Download the completed video

curl https://api.quotaflow.ai/openai/v1/videos/$TASK_ID/content \
  -H "Authorization: Bearer $QUOTAFLOW_API_KEY" \
  --output seedance.mp4

You may instead fetch the completed response's temporary data[0].url directly without an Authorization header. It expires after 15 minutes; poll the task again to refresh it.

Treat Seedance tasks as async media jobs. Portrait, identity, audio-reference, 1080p, 4k, or longer-duration jobs can take longer and may fail model safety review. Store task ids, poll idempotently, and show a pending state in your UI.

Notes

• Use stable public HTTPS URLs for reference media and keep them reachable until the task finishes.
• Prefer arrays (image_urls or reference_images) even for one image so your client is ready for richer reference workflows.
• If the selected key does not list seedance-2.0-fast from /models, do not retry as a transient error; request Seedance video access for that key. Confirm your package supports 4k before exposing a 4k selector to end users. 4K jobs may be slower than 720p fast jobs.
• If a portrait or person-reference job fails safety review, adjust the prompt/reference media and submit a new task.

Next steps

• Read the Videos endpoint guide.
• Check /models with your production key for currently enabled video models.
• Browse all supported models.