Skip to main content

Voice and audio

Updated October 10, 2026

Audio in InfluenceBot comes from your persona's voice. You pick the voice once on the persona, then add it to scenes when you render a storyboard. The editor mixes the audio and sets the loudness when it renders the final video.

1. Give the persona a voice

On the persona's Voice tab, pick a voice, listen to a sample and select Use this voice. See Pick a voice for your persona.

2. Add a voice-over to the scenes

  1. Open the storyboard and go to Render storyboard ✦.
  2. With Preview → video, tick Add voice-over. Each scene's Spoken text is read in the persona's voice.
  3. Or choose Talking avatar: each scene with spoken text becomes a video of the persona's avatar speaking.
  4. Select Estimate cost, then confirm.

Voice-overs are linked to their scene videos (lineage Voice-over in the asset library). When the storyboard finishes, the timeline in the Editor already includes them.

3. Render with the right loudness

When you render in the editor, the audio is mixed and normalised to the standard short-video loudness. The render card shows the output format, for example "1080×1920 · 30 fps · H.264/AAC · loudness −14 LUFS". During rendering you may see Mixing audio and Normalizing audio.

Captions from the voice

Transcribe captions ✦ in the editor turns the voice-over (or the video's own sound) into timed captions. See Add captions.

What isn't available yet

  • The editor has no controls yet to add, move or adjust music or other audio tracks. The storyboard's Music field is a note for planning.
  • If a timeline contains music without a clear commercial licence, the editor shows "Music license needs review".
  • Audio files uploaded to Assets are stored and can be downloaded, but they can't be placed on a timeline from the editor yet.

While voice providers run in MOCK mode, voices and voice-overs are simulated and labelled MOCK.

What it costs

Voice-overs, talking-avatar scenes and the render are paid. You see the estimate first; when you confirm, up to that amount is reserved and only the actual cost is charged. See How credits work.

What can block you

  • A cloned voice needs an active voice consent. See Consent for real people and voices.
  • No voice chosen: voice-overs can't use the persona's voice until you pick one. The Voice tab says "No voice chosen yet."
  • An error on a line of the estimate means that step can't run as configured. Fix the cause before you confirm.

More in Creating content