What is Baby Lip Sync?

Where we found it

The post we follow went up on TikTok on 29 September 2026, runs 44s and had about 4.8 million plays on 12 November. TikTok shows no AI label on it. The earliest post of the idea that we found dates from 24 July and had around 384,000 plays.

We follow roughly the first 12s of the post, sped up to 10s. The people and the room in it never appear in your video, and the song is not part of it.

The clip, beat by beat

• 0-5s: the first singer fills the frame and mimes straight into the lens, with huge faces and sweeping gestures.
• 5-10s: the lens glides to the second singer, who takes the chorus over with the same energy while the first one cracks up at the edge.
• With a third photo: the beats shorten to about 3s each, and the camera swings on to the third singer for the finish.

Why it catches on

Everyone knows the feeling of singing a chorus at a friend and handing over the phone. The camera pass turns it into a relay, and the goofier the faces, the better it plays.

In your version the place, the outfits and the light are generated, while every singer keeps the face or fur from their own photo.

Give it a go

You can jump straight into the generator or skim the how-to guide beforehand. For a group of friends dancing instead see Babydoll Dance.