Lip-sync a clip to your audio
Sometimes the voice is the asset: your recorded hook, a client's approved voiceover, the chorus of a track. Reference audio lets the clip follow it — upload up to fifteen seconds and the render times mouth and motion to the sound instead of generating a voice.
It works in Text to Video on Draft, Standard and Pro, with a presenter you describe; the People tier and AI influencer clips skip reference audio. Keep the audio clean and the prompt about the setting and the framing; the audio carries the performance.
Starter prompt
“Close framing on a presenter at a bright kitchen counter, looking into the lens, handheld feel”
What this solves
- A recorded voiceover that needs a face
- A hook you said better on a voice memo than on camera
- A described singer on the chorus of your track
- Client-approved audio that cannot change
The steps
Trim the audio to 15 seconds or less
The reference audio slot accepts clips up to 15 seconds; the video length should match.
Text to VideoDescribe the presenter and the framing
Close framing on the face makes the sync obvious; leave the words to the audio.
Text to VideoScore the rest of the video
A 30-second bed from Music & Soundtracks sits under the synced clip for 2 credits.
Music & SoundtracksQuestions
What audio formats work?
MP3, WAV and M4A up to 15 seconds. The Studio checks the length before upload.
Should I still write the line in the prompt?
No — with reference audio attached, describe only the scene and the framing. The audio drives the performance.
Can I use a song?
Only music you have the rights to. A track from Music & Soundtracks is yours to use.
Lip-sync a clip to your audio
Open with these settings