GitLab Transcend - Intelligent orchestration
Mini toolsSync 3
By sync.
Sync-3 is a studio-focused video lipsync model designed for high-quality dubbing and localization across many types of content, including films, podcasts, games, animation, and other “in the wild” footage. Sync describes it as its most intelligent lipsyncing model, approaching the task with wide spatial context so it understands the scene rather than only the mouth region. The public page highlights 4K 60 FPS support and emphasizes strong performance on sharp angles, side faces, close facial detail, low lighting, multiple speakers, and preservation of acting and emotion across languages, with the goal of reducing retakes and manual fixes.
New
Multimodal
Gen 3
Overview
Sync-3 is Sync.’s advanced lipsync and visual dubbing model for video localization. It is built to generate near-indistinguishable mouth-sync across difficult real-world footage, using broader spatial scene understanding to handle side profiles, angle changes, multiple speakers, close-ups, low light, and emotion preservation.
About sync.
sync. (also known as sync.labs) builds generative AI video models to modify and synthesize humans in video. Its flagship product is a state-of-the-art zero-shot lip-syncing model that synchronizes any video to any audio in any language in near real-time at up to 4K resolution, enabling visual dubbing, automated dialogue replacement, and multilingual content localization.
Industry:
Software Development
Company Size:
6
Location:
San Francisco, California, US
Website:
sync.so
Tools using Sync 3
No tools found for this model yet.
Last updated: April 7, 2026
