Jimeng Video

Tianshu offers Jimeng Seedance text-to-video via the jimeng-video group, in two tiers: video-ds-2.0 (standard) and video-ds-2.0-fast (fast). Billed per clip, any length.

Option 1: Playground (recommended, easiest)

1. Open Console → Playground and switch to the Video tab at the top.
2. Pick group jimeng-video and model video-ds-2.0 or video-ds-2.0-fast.
3. Enter a scene description, set aspect ratio / duration as needed, and click Generate video; ready in ~2 min, play/download inline.

Option 2: API

Jimeng video is an async task: submit to get a task_id, then poll for the result. Use a token from the jimeng-video group.

curl 'https://api.tian-shu.org/v1/video/generations' \
  -H 'Authorization: Bearer your jimeng-video group token' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "video-ds-2.0",
    "prompt": "an orange cat napping in the sun, cinematic",
    "seconds": "5",
    "aspect_ratio": "16:9"
  }'

After submitting you get a task_id; poll GET /v1/video/generations/{task_id} until status becomes completed, then fetch the finished video.

Available models and per-clip prices follow the jimeng-video group on the Model Marketplace.