seedance-2.0-fast-multi API
Seedance 2.0 has 18 public models: Standard, Fast and Mini tiers, each with text-to-video (t2v), first/last-frame image-to-video (i2v) and mixed image, video and audio references (multi). IDs without global run in China; IDs with global- are the international versions.
Capabilities
- Model
- Video
- API model
seedance-2.0-fast-multi- Endpoints
- POST /v1/videos, GET /v1/videos/{id}
- Parameters
- prompt (required), seconds: 4–15, metadata.duration: -1 (auto), metadata.resolution: 480p / 720p / 1080p / 2k / 4k (Standard: native1080p / native4k), metadata.ratio: adaptive / 16:9 / 4:3 / 1:1 / 3:4 / 9:16 / 21:9, metadata.seed, metadata.generate_audio, images: 1–2 (i2v), metadata.content: ≤9 images, ≤3 videos, ≤3 audio (multi)
Current pricing
≈¥1.35–10.82 per request
Gallery examples
GardenAPI model:
seedance-2.0-global-fast-multiGarden — explore this seedance-2.0-global-fast-multi creation on OpenZhenZhen, with its original prompt and a reusable generation example.
Yellow Dress on the Sunset CourtAPI model:
seedance-2.0Yellow Dress on the Sunset Court — explore this seedance-2.0 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
Flying Side Kick in Red-Gold Muay Thai ShortsAPI model:
seedance-2.0Flying Side Kick in Red-Gold Muay Thai Shorts — explore this seedance-2.0 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
She Kissed Me Before the NightAPI model:
seedance-2.0-global-fast-multiShe Kissed Me Before the Night — explore this seedance-2.0-global-fast-multi creation on OpenZhenZhen, with its original prompt and a reusable generation example.
Quick start
Example workflow: draft a 5-second clip with t2v from a prompt alone; switch to i2v and pass 1–2 images (30 MB each at most) to fix the first frame or the first and last frames; use multi to follow a reference video, swap a person or add sound, listing images, videos and audio in metadata.content and naming them in the prompt as @Image 1 and @Video 1. Every model requires a prompt. Duration is 4–15 seconds; send metadata.duration: -1 for automatic length. Multi accepts up to 9 images, 3 videos and 3 audio clips; a reference video switches the task to the with-video rate and its length is billed. 1080p, 2k and 4k are upscaled tiers with a per-second surcharge; Standard also offers native1080p and native4k. Real-person mode is always on for i2v and multi.
cURL
curl --fail-with-body https://www.zzai.app/v1/videos \
-H "Authorization: Bearer $SEEDANCE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"metadata":{"content":[{"image_url":{"url":"https://your-cdn.example/person.png"},"type":"image_url"},{"type":"video_url","video_url":{"url":"https://your-cdn.example/source.mp4"}}],"resolution":"720p"},"model":"seedance-2.0-fast-multi","prompt":"Replace the person in @Video 1 with the person in @Image 1, keep the camera motion.","seconds":"5"}'Python
import json, os, requests
payload = json.loads(r'''{"metadata":{"content":[{"image_url":{"url":"https://your-cdn.example/person.png"},"type":"image_url"},{"type":"video_url","video_url":{"url":"https://your-cdn.example/source.mp4"}}],"resolution":"720p"},"model":"seedance-2.0-fast-multi","prompt":"Replace the person in @Video 1 with the person in @Image 1, keep the camera motion.","seconds":"5"}''')
response = requests.post(
"https://www.zzai.app/v1/videos",
headers={"Authorization": "Bearer " + os.environ["SEEDANCE_API_KEY"]},
json=payload, timeout=120,
)
response.raise_for_status()
print(response.json()["id"])JavaScript
const response = await fetch("https://www.zzai.app/v1/videos", {
method: 'POST',
headers: { Authorization: 'Bearer ' + process.env.SEEDANCE_API_KEY, 'Content-Type': 'application/json' },
body: JSON.stringify({"metadata":{"content":[{"image_url":{"url":"https://your-cdn.example/person.png"},"type":"image_url"},{"type":"video_url","video_url":{"url":"https://your-cdn.example/source.mp4"}}],"resolution":"720p"},"model":"seedance-2.0-fast-multi","prompt":"Replace the person in @Video 1 with the person in @Image 1, keep the camera motion.","seconds":"5"}),
});
if (!response.ok) throw new Error(await response.text());
console.log((await response.json()).id);Save the returned id as TASK_ID in the query example. Query every 3–5 seconds until success or failure; submission does not mean generation is complete. Save result files promptly. status: completed / failed; metadata.url.
curl --fail-with-body https://www.zzai.app/v1/videos/$TASK_ID \
-H "Authorization: Bearer $SEEDANCE_API_KEY"Frequently asked questions
How do I call this model?
Create an API key, choose the model name below, and send a request to the documented OpenAI-compatible endpoint. Final billing follows the current price and usage parameters.