Skip to main content
POST
Create video generation

Introduction

Submit an async Seedance job; returns task_id for polling. Provide at least one of content or prompt; content is recommended.

Authentication

string
required
Bearer Token, e.g. Bearer sk-xxxxxxxxxx
string
application/json

Request body

string
required
seedance-2-0 / seedance-2-5 / seedance-2-0-mini / seedance-2-0-fast. seedance-2-5 differs from 2.0 in duration, reference-asset limits, and video editing/extension โ€” see Seedance 2.5 Notes
array
Multimodal array (see below); at least one of content or prompt required
string
Text prompt; simplified alternative to content
integer
default:"5"
-1 (auto) or 4โ€“15 seconds; seedance-2-5 supports up to 30 seconds โ€” see Seedance 2.5 Notes
string
default:"720p"
480p, 720p, 1080p, 4K (seedance-2-0-fast and seedance-2-0-mini support only 480p / 720p)
string
default:"16:9"
16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive (recommended with first-frame images; not available in reference-image/video mode)
boolean
default:"true"
Auto-generate audio; set false when using reference_audio
boolean
default:"false"
Add watermark
integer
default:"-1"
Random seed; -1 for random; fixed positive integer for reproducibility

content array

Rules: first_frame and reference_image are mutually exclusive; last_frame requires first_frame.
asset:// must match asset_type: Image โ†’ image_url; Video โ†’ video_url; Audio โ†’ audio_url. Putting audio in image_url returns InvalidParameter.

Direct URL limits in content

When passing http(s) URLs directly in content (not asset://), resources must meet these limits. asset:// references share the same numeric limits; asset library images allow more formatsโ€”see Create asset.

Examples

Text-to-video:
Simplified prompt (text-to-video):
Image-to-video (first frame):
Image-to-video (first and last frame):
Asset library + multimodal:
Only active assets may be used; all referenced assets must belong to the same asset group.

Response

Next: Get video generation