How to Make a Lip Sync Video with AI

1. Add a product photo, or use the Write my own prompt tab if there is no product. 2. Press Choose your creator and pick a ready-made face, one of your saved faces, or upload your own photo. 3. Press Write my ad — free; the storyboard gives each beat an action and a spoken line. 4. Rewrite the spoken lines in your own words; they are what the person says on camera. 5. Pick 9:16 or 16:9 and render.

Keep each spoken line to one breath. Short lines are easier for any model to voice cleanly inside a 15-second clip.

If the voice or song already exists, the Cinema studio takes it as reference audio. Seedance 2.5 accepts up to 10 MP3 or WAV files and can start from audio alone; Wan 3.0 accepts up to 5 audio files together with a photo or clip of the person. A 5-second Seedance 2.5 clip costs 115 credits (~$1.43).

Reference audio guides the clip rather than locking every syllable to a frame, so watch the result before you post it, and render a second version if a line needs tighter timing.

The model list marks silent models with a No sound badge, so you can see before paying whether a clip will come back mute. Kling's own page says Kling VIDEO 3.0 generates dialogue together with lip movements in one pass; Kling V3 is in the Cinema studio at 50 credits (~$0.62) for 5 seconds.

A 15-second talking clip from the UGC studio costs 278 credits (~$3.47) at 720p, and the storyboard with its spoken lines is written free. Sign-up gives 100 free credits with no card, and the exact charge shows before anything renders.