Kling v2.6 Motion Control

Kling v2.6 Motion Control is a video generation model built around motion transfer rather than free-form synthesis. It takes a short reference clip, between three and thirty seconds long, and applies its gestures, facial expressions, lip-sync, and camera movement to a newly generated scene with frame-accurate fidelity. The workflow on Vercel AI Gateway is exposed as a text-prompt-driven video endpoint, where a description of the desired scene is combined with the uploaded reference assets so the system can faithfully replay the reference motion while reshaping the surrounding context.

Accessed through the Vercel AI SDK, the model is invoked via the experimental video generation function with the identifier under the Kling AI provider, making it straightforward to drop into a TypeScript or JavaScript application that already uses Vercel's AI primitives. The practical strength of the model lies in producing videos where a character moves convincingly according to a known choreography or performance, which is useful for content that needs consistent, repeatable motion such as dance sequences, lip-synced dialogue, and stylized camera paths. It fits workflows that already have a recorded motion source and need a flexible target scene, rather than fully open-ended text-to-video generation.