Why this matters

Saved narration could appear in a project while its playable audio stayed unavailable, leaving preview and MP3 or MP4 export blocked. Narration restore now keeps the account session attached, loads only the byte ranges the player needs, and automatically recovers the durable audio into a local playback copy if the stream cannot continue.

What changed

Reopening a project now reconnects its saved narration with authenticated streaming, responsive seeking, and an automatic local recovery path when streaming is interrupted.

More expressive narration

Choose from 30 familiar voices with gender and tone shown in the picker. Delivery controls guide the performance, while pauses, laughs, sighs and breaths can be placed in individual audio scenes. Your spoken words stay intact, and these cues stay out of subtitles. New generations use Gemini 3.8 Flash; saved audio stays playable.

How to use it

  1. Choose a voice

    Open Narration audio and choose a voice, using the gender or Favorites filters if useful.

  2. Shape the delivery

    Choose a Delivery preset and adjust Custom direction for pronunciation, accent or emotion.

  3. Place voice cues

    Expand Pauses & vocal effects, choose an Audio scene, and insert a cue in Narration with cues.

  4. Generate narration

    Choose Generate MP3 to create the narration with the selected voice and performance.

What has changed over time

  1. Improved

    Give narration more expressive delivery

    Narration audio now uses Gemini 3.8 Flash, with Storytelling, Excited and Whisper delivery plus precise pauses and vocal cues.

  2. Fixed

    Saved narration now comes back reliably

    Reopening a project now reconnects its saved narration with authenticated streaming, responsive seeking, and an automatic local recovery path when streaming is interrupted.