Seedance Lip Sync: Turn a Photo and a Voice Into a Talking Video

Seedance 2.0 accepts reference audio alongside your images. When the prompt tells it to use that track as the voice (the @audio1 mention), the model generates the face speaking or singing those exact sounds, mouth shapes matched to the words. That is Seedance lip sync in one sentence: your photo, your audio, one render.

It works for more than dialogue. A song from Suno, a recorded voiceover, a podcast clip: feed any of them in and Seedance lip sync makes the character perform it. Each generation takes up to 9 reference images and 3 audio tracks, so you can lock the face and the voice in the same request.

Everything runs on this page. Pick a Seedance lip sync template above, add your files, and the clip renders in the browser. No editing timeline, no separate sync pass, no software to install.

Short lines sync better than long speeches; five to ten words per sentence gives the model clear anchors. If a Seedance lip sync take drifts, trim the audio into smaller pieces and render them separately rather than re-rolling the same long clip.

Name things explicitly in the prompt. "The person in @image1 speaks the dialogue from @audio1" beats a vague "make it talk" — the mention tells the model exactly which file drives the mouth.