AI Lip Sync From a Photo: Make Your Photo Sing Your Song

1. Add a photo of the singer on the left and of the partner on the right, or one photo of you both. A pet can take the partner’s spot.
2. Tap Change on the sound row (the Soundtrack & moves drawer) and choose Your song. The model stays on Wan 3.0, the one that plays your audio.
3. Upload an MP3, WAV or M4A (a phone voice memo is M4A). Move the slider to the 2 to 15 seconds you want.
4. Pick the length and format, check the credits on the button and tap Create video. It takes about 5 to 15 minutes.

• Clear vocals with little background noise. A voice memo recorded close to the phone works.
• A part that starts on a word, not on silence or a long intro.
• Rap, singing and speech all work; fast rap is the hardest to follow.
• Only use audio you have the right to use. For the official “Hotel Lobby” recording, add the licensed sound inside TikTok or Reels when you post.

• Only the left person lip-syncs your audio. Both singing your song in turns is not possible: two people, one song explains the workarounds.
• There is no single-person mode: the video is always two performers.
• The audio is 15 seconds at most per video.
• It does not clone a voice. The sound is your file, as you uploaded it.

For two voices trading lines, pick AI rap on a Seedance model with Both rap on. The AI writes the verse; your audio is not used.

This is one of our own renders, on Wan 3.0 with the built-in Hotel Lobby track instead of an upload: the left performer mouths the words into the mic. We have not published a render with someone’s own song, so judge the mouth movement here, not the audio.

Our render