Seedance 2.5
In Omnibo, from 123.6 cr/s.

You give it

• A text prompt
• A start frame and an end frame
• Up to 30 reference images
• Up to 10 reference videos
• Up to 10 reference audio files
• A clip to edit
• A clip to extend
• A clip to upscale

You get back

• A clip of 4 to 30 seconds
• In 480p, 720p, and 1080p
• In these shapes: 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9
• With sound you can switch off
• Some resolutions can come back as HEVC video, which Omnibo transcodes to H.264 for playback

In Omnibo

1. Open Video in the app and pick Seedance 2.5 in the model selector.
2. Choose a length from 4 to 30 seconds and a resolution: 480p, 720p, and 1080p.
3. Add a start frame or an end frame to pin how the clip opens or closes.
4. Attach references to steer it: images, videos, and audio.
5. To work on a clip you already have: edit, extend, and upscale.

Try Seedance 2.5 →

Over MCP

Connect Omnibo to your AI assistant over MCP, then call these tools with the model id seedance-2-5.

• list_models
• balance
• generate_video
• edit_video
• extend_video
• upscale_video
• job_status
• wait_for_job

An example call to generate_video

{
  "model": "seedance-2-5",
  "prompt": "A slow push-in on a lighthouse at dusk, rain on the lens",
  "duration_seconds": 10,
  "resolution": "720p"
}

How to connect over MCP