sync.

sync. provides AI lip syncing and visual dubbing through a browser Studio and developer API. Its introduction describes supplying a video and audio track so the resulting mouth movements match the replacement speech. The useful distinction is between preparing the new voice track and adapting the on-screen performance to it.

Prepare the recording and replacement audio

The documented workflow accepts common video and audio formats, including MP4 video and WAV or MP3 audio. Inputs can be supplied through supported upload and URL routes. Start with material you are authorized to process and a replacement track whose timing and pronunciation you have reviewed.

For a dubbed narration pipeline, the documentation explicitly names ElevenLabs as a text-to-speech service that can be combined with sync. That makes voice generation a separate stage from lip syncing. The built-in dubbing flow supports specified target languages; providing your own translated audio is a different route.

The publisher illustrates changing an English-speaking recording to Japanese speech. That example explains the intended workflow, but it is not a promise that every face, camera angle or translation will produce an acceptable final shot.

Compare models in the browser

Studio is the web application for exploring and comparing models. The documentation describes multiple models with different quality and speed tradeoffs, so use a representative clip before selecting one for a longer production.

The product site also describes uploading a video and supplying cloned, generated or original audio. Its claims about difficult lighting, obstruction and camera movements are product positioning, not measurements this directory has reproduced.

For an evaluation, inspect facial edges, mouth timing and transitions against the original performance. Keep the original recording and approved voice track available so an unsatisfactory generation does not become the only version of the scene.

Integrate without assuming unlimited usage

The API has official Python and TypeScript SDKs. Webhooks can report asynchronous status, while batch processing is subject to plan access and limits. These are developer capabilities rather than an assurance that the browser and API have identical allowances.

The documentation points to a cost-estimation endpoint and the account billing dashboard for current pricing and usage. Check the selected model, duration and plan before submitting a batch; published example rates are not fixed quotations for future work.

sync. is best evaluated as one part of a dubbing or localization pipeline. It does not remove the need to review the translated script, obtain appropriate permissions or check the finished video before release.