Lip sync a character to a voice line

Last updated: October 12, 2026

Make a character speak a voice line: turn a portrait into a talking clip, or match the lips in an existing video to new audio. Use it for dialogue in cutscenes and trailers, a talking-head dialogue box, or replacing a line after the video is made.

Before you start

  • Have the voice line as an audio file in your library. To make one, see Create speech and voice-over.
  • Have the character as a still image (Lip sync image) or as a video (Lip sync video). The face should be clearly visible and turned mostly towards the camera.
  • Models limit how long the audio can be, and the clip length follows the audio. Keep each line short; split long dialogue into several runs.
  • Each run costs Creative Units. The cost shows on the Lip sync button before you run. See Creative Units for how it is charged.

Steps

  1. On the session's board, hover over the portrait (or the video) and click View to open it full-size.

  2. In the tools beside the viewer, click Lip sync image for a still, or Lip sync video for a video. The form opens with the file you opened as its source, on the default model.

  3. Click Audio, pick the voice line in the picker, and confirm with Select 1 asset. You can also drop an audio file onto the slot.

  4. Click Lip sync.

  5. Wait while the clip generates. The viewer shows the run's progress, and you can keep working while it runs.

Result

A video of the character speaking the line, with the audio in the file. It joins the version strip on the left of the viewer, and the viewer plays it when it finishes; the original is unchanged. It also lands on the session's board. To save it, click Download under Export in the viewer, or right-click its tile and choose Download.

Production tips

  • Start from a clean face. Front or three-quarter view, mouth closed or neutral, nothing in front of the lips, even light. Profiles, masks, beards that hide the mouth and fast head turns sync poorly.
  • Use clean audio. A dry voice with no music or effects underneath syncs best. Run Isolate voice or Clean audio on a recording first, and add the music and sound effects back in your edit.
  • Check the hard consonants. Play the result with sound and watch the mouth on p, b and m. If they slip, try a cleaner recording or a shorter line.
  • Animate first, then sync. For a character who should also move, make the clip with Create video from your approved still, then run Lip sync video on it. Ask for a calm head and a closed mouth in that clip so the sync has room to work.
  • Line by line. Sync each line of dialogue as its own run and cut them together. A retake then costs one line, not the scene.
  • Localise the same shot. Keep the portrait or video and run it once per language with each translated voice line. To translate the speech in a finished clip, use Dub video instead.
  • A whole cast of talking portraits. See Turn an expression sheet into talking portraits for expressions, lip sync and engine hand-off together.

If something goes wrong

  • The lips drift off the words. Use a cleaner voice line, check that the face stays visible for the whole clip, and try another model from the model picker.
  • The run will not start because the audio is too long. Trim the line or split it into shorter runs.
  • The face warps or the head drifts out of frame. Start from a front-facing head-and-shoulders portrait where the face fills more of the frame.
  • The generation failed. The viewer and the tile on the board show the failure. Run it again; if it fails again, change the model. See Fix a failed or stuck generation.

Related