How to Install / Upgrade: Seedance 2.5 now available on Vercel AI Gateway

Seedance 2.5 now available on Vercel AI Gateway

Seedance 2.5 from ByteDance is now available on Vercel AI Gateway. It can generate up to 30 seconds in a single clip while holding camera movement and continuity without stitching shots together in post. Short clips can be extended with character, scene, and camera movement carried over. Seedance 2.5 supports text, video, image, and audio as inputs in the same request, following a subject's appearance from one asset and the motion or related cues from others as described in the release summary.

To use it, route generation through Vercel AI Gateway and select Seedance 2.5 as the model. Send your request with the inputs you need in one call—text, video, image, and/or audio. For a new clip, provide the prompts and reference assets that define the subject and motion. To extend an existing short clip, pass that clip back through the gateway with Seedance 2.5 so character, scene, and camera movement continue from the prior footage rather than starting a separate shot.

Watch for multi-input requests: the model is meant to take text, video, image, and audio together, so omit only inputs you do not need and do not assume a text-only path is required. Continuity depends on extending a clip through the model instead of stitching separate clips in post; if you need a longer sequence with matching character, scene, and camera movement, extend the short clip rather than generating disconnected segments. Verify success by confirming the response comes from Seedance 2.5 via AI Gateway, that a single clip can reach up to 30 seconds without post stitching for camera continuity, and that an extended clip keeps the same character, scene, and camera movement from the source.