Quickstart

Generate your first AI video with the Seedance 3.0 API — create an API key, start a generation, and poll until the video is ready.

The Seedance 3.0 API gives you programmatic access to the same video, image and music models as the website generator — Seedance 2.5, Kling, Wan, HappyHorse, LTX and more — billed from the same credit balance.

Every generation is asynchronous: you create a job, get back an id, then poll that id until status is success and the output URLs are ready.

Before you start

Create a key on the API Keys page and keep it secret. Generations spend credits from your account, so top up on Pricing if your balance is low.

Base URL

https://seedance3ai.video/api/v1
MethodEndpointPurpose
GET/modelsList the models you can call
POST/generationsStart a video, image or music job
GET/generations/{id}Poll a job's status and get its outputs
GET/creditsCheck your remaining credits

1. Start a generation

Shell
curl https://seedance3ai.video/api/v1/generations \
  -H "Authorization: Bearer $SEEDANCE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance25-text-to-video",
    "prompt": "A red fox running through fresh snow at sunrise, cinematic tracking shot",
    "options": {
      "duration": "5",
      "resolution": "720p",
      "aspect_ratio": "16:9"
    }
  }'

The response is a generation object with status pending or processing:

{
  "id": "0b6f4c1e-8f1a-4d0e-9a51-1d3c2a7e9b10",
  "object": "generation",
  "status": "processing",
  "model": "seedance25-text-to-video",
  "model_name": "Seedance 2.5",
  "media_type": "video",
  "scene": "text-to-video",
  "prompt": "A red fox running through fresh snow at sunrise, cinematic tracking shot",
  "outputs": [],
  "error": null
}

2. Poll until it finishes

Shell
curl https://seedance3ai.video/api/v1/generations/0b6f4c1e-8f1a-4d0e-9a51-1d3c2a7e9b10 \
  -H "Authorization: Bearer $SEEDANCE_API_KEY"

When status becomes success, outputs holds the result URLs:

{
  "id": "0b6f4c1e-8f1a-4d0e-9a51-1d3c2a7e9b10",
  "object": "generation",
  "status": "success",
  "outputs": [
    { "type": "video", "url": "https://.../fox.mp4" }
  ],
  "error": null
}

Video jobs usually take 1–5 minutes. Poll every 5–10 seconds.

Full example

Python
import os
import time

import requests

API = "https://seedance3ai.video/api/v1"
HEADERS = {"Authorization": f"Bearer {os.environ['SEEDANCE_API_KEY']}"}

job = requests.post(
    f"{API}/generations",
    headers=HEADERS,
    json={
        "model": "seedance25-text-to-video",
        "prompt": "A red fox running through fresh snow at sunrise",
        "options": {"duration": "5", "resolution": "720p", "aspect_ratio": "16:9"},
    },
).json()

while job["status"] in ("pending", "processing"):
    time.sleep(8)
    job = requests.get(f"{API}/generations/{job['id']}", headers=HEADERS).json()

if job["status"] == "success":
    print(job["outputs"][0]["url"])
else:
    print("failed:", job["error"])
TypeScript
const API = 'https://seedance3ai.video/api/v1';
const headers = {
  Authorization: `Bearer ${process.env.SEEDANCE_API_KEY}`,
  'Content-Type': 'application/json',
};

let job = await fetch(`${API}/generations`, {
  method: 'POST',
  headers,
  body: JSON.stringify({
    model: 'seedance25-text-to-video',
    prompt: 'A red fox running through fresh snow at sunrise',
    options: { duration: '5', resolution: '720p', aspect_ratio: '16:9' },
  }),
}).then((res) => res.json());

while (job.status === 'pending' || job.status === 'processing') {
  await new Promise((resolve) => setTimeout(resolve, 8000));
  job = await fetch(`${API}/generations/${job.id}`, { headers }).then((res) =>
    res.json()
  );
}

console.log(job.status === 'success' ? job.outputs[0].url : job.error);

Next steps

  • Authentication — how keys work and how to keep them safe
  • Models — model ids, scenes and accepted options
  • Generations — every request field, image-to-video and reference inputs
  • Errors — status codes and retry guidance