AI voiceover & lip sync

Turn a script into voiceover on the audio track, then generate lip-sync keyframes for a character from the same clip. When the copy changes, only the text changes.

What it does

• Voiceover from a script: the AI button in the Audio panel header opens the voiceover panel; AI+ → AI voiceover (TTS) in the top bar opens the same panel. Enter the script, describe the voice and adjust speed.
• Automatic placement: the generated audio is added to the project and dropped on the current scene's default audio track. The playhead jumps to the clip start.
• Per-clip voiceover tab: selecting a TTS clip adds a Voiceover tab to the dock, where you can edit the script and voice, regenerate, or generate subtitles in one click.
• Duration fitting: Fit the scene length to the voice duration is on by default, so the scene grows or shrinks with the voice.
• Visible usage: the panel accumulates call count, audio duration, file size and token usage.
• Lip sync: select a character layer, choose a scene audio clip in the Lip sync panel, enter the script, and generate keys from the waveform and syllables.
• Mouth set (10): bind a mouth texture to each of M F T L CH A E I O U. M is the resting shape, used for still and closed poses, so there is no separate base image. Unconfigured shapes fall back to the built-in default mouth set.
• Mouth slots (multi-mouth): one character can carry several mouth slots, and lip-sync sequences are generated per slot. Slots are added in the Character Template editor.

How to use

1. Click the AI button in the Audio panel header (or AI+ in the top bar) and enter the script in the voiceover panel, then choose a voice.
2. Click Generate voiceover. The audio appears on the current scene's default audio track.
3. Select the character layer, switch the dock to Lip sync, and choose that clip under Audio.
4. Enter the script text. Without text, lip sync cannot align to syllables and generation is rejected.
5. Click Generate lip sync, then choose a mouth slot and fill in the 10-cell mouth set, or match textures by file name.

::: note Where the voice comes from Voiceover requests are proxied by the server, and the browser never sees the credentials. Without separate TTS credentials, the project generates placeholder audio to exercise the full generate → place → export path. :::

Boundaries

• Lip sync follows the clip start, offset and speed. With no audio selected, it fills the layer window.
• The mouth set only takes effect after a mouth slot is selected. Without one, the procedural mouth remains in charge; mouth slots are added in the Character Template editor.
• Voiceover generation makes real calls and consumes quota; the panel reports usage as it accumulates.

With sound in place, continue to Skeletal animation.