Wav2Lip

Matching mouth movements to a voice track by hand is slow, fiddly work, and Wav2Lip exists to take that job off your plate.

Wav2Lip is an AI lip sync generator that takes an audio file plus either a still image or existing video and produces a talking face video with the lips synced to the speech.

The current description mentions Deep SyncNet and GAN models behind the alignment, which is where its accuracy comes from.

Two uses stand out.

You can animate a static portrait so it appears to speak, which suits avatar videos and explainers.

You can also resync footage you already have to a new audio track, which is the core of AI dubbing, for example when a voiceover is replaced or translated.

Creators, educators and developers are the obvious audience.

Wav2Lip works with any audio input, so you are not tied to a particular voice source.

Expect to judge the results on your own clips, since lip sync quality varies with the footage.