sync. x Kapwing: powering lipsync for everyone
sync. is now powering Kapwing's AI Personas, letting Kapwing's millions of creators make zero-shot, talking-head videos from just 15 seconds of footage.
The bet at sync. has always been the same: build one general ai lip sync model that works on any face, no pre-training. We’re the team behind Wav2Lip, the first zero-shot lipsync model, open-sourced four years ago. Plenty of models have shipped since, but most of them still need training footage. With sync., you don’t.
Today we’re shipping that capability inside Kapwing. Their new AI Personas feature is powered by sync., and it lets users create talking-head videos from just 15 seconds of footage, no avatars, no cloning queue, no training. It’s the easiest way to make polished, multilingual video at scale that we’ve seen anyone build, and it’s available to Kapwing’s millions of creators today.
Working together with Kapwing
Kapwing is one of the best web-based video editors out there. Trim, resize, subtitle, all the small tools that turn raw footage into something postable, all in the browser. Their workflows are smooth in a way that most editors aren’t, which is the kind of thing you only learn by spending a real amount of time inside them.
We’ve wanted a partner like Kapwing for a while. The thing that gets a researcher out of bed isn’t the model itself, it’s seeing the model used by a million strangers. The moment we saw how Kapwing’s users felt about the editor, putting sync. inside it was an obvious fit.
What’s happened since has been the best version of a partnership. Kapwing experiments fast, gives us real feedback, and ships things that meaningfully change the model in the process.
When they came to us with “AI Personas,” it was clear they understood the leverage of zero-shot. While other tools were still asking users to upload training footage and wait hours or days, Kapwing built something that just works off the shelf, because the lip sync ai underneath doesn’t need training to begin with.
How sync. works in Kapwing
With sync. inside Kapwing’s video editor, you can create personas from your own clips and lipsync them right there. Practically:
- AI Personas, no more lining up voiceovers by hand. Pick a sync. model inside the editor, and Kapwing’s Personas feature handles the rest. Studio-quality lip sync in minutes.
- Global ready, built-in support for over 29 languages, so you can localize the same video for every market without re-recording. Visual dubbing without the dubbing studio.
The beginning of a great partnership
We’re excited to power Kapwing with sync. This is round one. A lot more is in the pipeline that should make video editing feel as fluid as editing a doc.
Excited to see what their users build with it.