Seedance 2.5

Seedance 2.5 creates cinematic AI videos from text prompts and reference materials, including images, video, and audio. Generate coherent 30-second clips in a single run, designed to deliver a complete story short or brand film without fragmented stitching.

The model focuses on shot-to-shot logic, smooth multi-shot transitions, and stable characters and lighting for consistent cinematic camerawork. You can reference motion, camera moves, transitions, character looks, scenes, and audio—and use real-person reference images to help maintain lifelike, consistent characters.

After generation, you can move into cloud editing to customize shot effects and packaging without third-party software. It also supports global multilingual workflows with accurate lip-sync across languages and localized on-screen text for cross-border content.

What can I generate with Seedance 2.5?

You can generate AI videos from text, images, video, and audio, and use references to guide motion, camera moves, transitions, character looks, scenes, and audio.

How long are the videos Seedance 2.5 can create?

Seedance 2.5 supports coherent videos up to 30 seconds, aimed at delivering a full story in one shot rather than stitched fragments.

Does Seedance 2.5 help maintain character and lighting consistency?

Yes—its improved shot-to-shot logic is designed to keep characters and lighting stable across multi-shot sequences for a more cinematic result.

Can I use real-person reference images?

Yes. You can reference real-person images to support lifelike, consistent characters across your generated video.

Is there built-in editing after the video is generated?

Yes. Seedance 2.5 includes cloud editing so you can customize shot effects and packaging without third-party software.

Does it support multilingual lip-sync and localized text?

Yes—Seedance 2.5 is designed for accurate lip-sync across languages and localized on-screen text for overseas and cross-border short videos.