AI Music Video Lip Sync

Two things are required: a photo of the person who should sing, and the audio track they should perform. A third, optional photo sets the scene — a stage, a bedroom, a street corner — and the singer is placed there instead of a generic set. Seedance 2.5 reads these as positional references, so the face from your portrait and the location from your scene photo are both carried into the video.

The model takes the audio itself as a reference rather than working from a text description, so mouth shapes follow the vocal frame by frame, and head and shoulder movement follows the beat. A clear, front-facing portrait and a clean vocal give the tightest result; heavy reverb, loud crowd noise, a face turned away from camera or a mouth covered by a mask or hand will loosen it.

Use music you already own or have licensed, or write one here first: Suno V6 at /models/suno-ai-v6 generates a full track from a text prompt, then you download the audio and upload it into this template. The two tools are built to be used back to back — write the song on the model page, perform it on this one.

Each audio file must be between 1.8 and 30.2 seconds and under 15 MB. The video itself renders 4 to 15 seconds, so trim to the section you actually want performed — usually the hook or the opening line of the chorus. If your track is longer, cut it before uploading rather than expecting the model to pick a highlight.

You are billed only for the generated seconds: 48 credits per second at 720p and 23 credits per second at 480p. So a 5-second 720p clip is 240 credits, the same clip at 480p is 115 credits, and a full 15 seconds at 720p is 720 credits. Reference images and audio files are not billed, so adding a scene photo costs nothing extra.

Only use music you own, music you have licensed, or music you generated yourself — for example with Suno V6. Uploading a commercial track you have no rights to can get the resulting video muted, blocked or taken down on TikTok, Reels, YouTube and elsewhere, and clearing the rights is your responsibility. The same applies to the person in the photo: get their permission before making them sing.