majiayu000 Seedance Video Generation
Generate AI dance and motion videos through AceDataCloud's Seedance (ByteDance) API.
Authentication
export ACEDATACLOUD_API_TOKEN="your-token-here"
Quick Start
curl -X POST https://api.acedata.cloud/seedance/videos \
-H "Authorization: Bearer $ACEDATACLOUD_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{"prompt": "a dancer performing contemporary ballet in a misty forest", "model": "seedance-1.0", "wait": true}'
Models
| Model | Best For |
|---|---|
seedance-1.0 |
General-purpose motion/dance video |
seedance-1.0-lite |
Faster, lighter generation |
seedance-1.0-pro |
Higher quality output |
seedance-1.5-pro |
Latest model, best quality |
seedance-acting-pro |
Character acting and expression |
Workflows
1. Text-to-Video
POST /seedance/videos
{
"prompt": "a street dancer doing breakdancing moves in an urban setting",
"model": "seedance-1.0-pro",
"resolution": "1080p",
"duration": 5,
"service_tier": "standard"
}
2. Image-to-Video
Animate a still image into a motion video.
POST /seedance/videos
{
"prompt": "the person starts dancing gracefully",
"image_url": "https://example.com/dancer.jpg",
"model": "seedance-1.5-pro",
"resolution": "720p",
"duration": 5
}
Parameters
| Parameter | Values | Description |
|---|---|---|
model |
See models table | Model to use |
resolution |
"360p", "540p", "720p", "1080p" |
Output resolution |
duration |
2 – 12 |
Duration in seconds |
service_tier |
"standard", "premium" |
Quality tier (premium = faster, higher priority) |
seed |
integer | Seed for reproducible results |
Task Polling
POST /seedance/tasks
{"task_id": "your-task-id"}
States: pending → succeeded or failed.
MCP Server
pip install mcp-seedance
Or hosted: https://seedance.mcp.acedata.cloud/mcp
Key tools: seedance_generate_video, seedance_generate_video_from_image
Gotchas
- Duration range is 2–12 seconds — values outside this range will fail
- Higher resolutions (1080p) combined with longer durations take significantly more time
premiumservice tier costs more but generates fasterseedance-acting-proexcels at facial expressions and character acting versus pure dance- Image-to-video requires a single
image_url— the person/subject in the image becomes the animated subject - Task states use
"succeeded"(not "completed") — check for this value when polling