sync-3

The tool provides AI‑driven lip‑sync and visual dubbing for video content, handling resolutions up to 4 K at 60 fps. It builds a spatial context of the scene, allowing it to align mouth movements with arbitrary audio while preserving acting performance, emotions, and details such as sharp angles, side faces, and close‑up lighting conditions. The model is designed to work across a wide range of source material, including movies, podcasts, games, and animations, and supports multiple speakers, low‑light environments, and extreme camera angles.

It targets creators, artists, and production studios that need to adapt existing footage to new languages or audio tracks without extensive manual re‑editing. By processing the entire visual context, the system aims to reduce retakes and manual fixes, delivering synchronized output that closely matches the original performance.

The software is offered as an API, allowing integration into existing pipelines. Documentation, changelogs, and example plugins are provided for developers, and the project is maintained by Synchronicity Labs. The current maturity level is experimental.