Generate audio
- Choose Text to voice
Select Text to voice in the desktop mode row, or open Add files & more and choose it from the mode selector.
- Enter the script
Paste the words you want spoken. The composer shows a character counter and accepts up to 40,000 characters. Check pronunciation, punctuation and spelling before sending.
- Choose a voice
Select a voice from the Voice control. If speed and style settings are available, choose the delivery you want before submitting.
- Send once and follow progress
Submit the script and wait for the audio card. Longer scripts may show progress across several parts. Do not submit another copy just because generation is still in progress.
Listen and download
When generation completes, use the audio player to listen and the download link to save the MP3. Check the whole recording before sharing it.
Guest audio is temporary in the current tab. Download it before closing or reloading. Signed-in audio cards show an expiry date when a saved file is available; download anything you need to keep before that date.
Revise or recover
If submission is rejected, correct the issue and retry the retained script. For an audio-card loading problem, Retry refreshes the file status; it does not rewrite your script.
When Edit script and regenerate is offered for a failed or expired recording, review the script and explicitly submit it again. Regeneration is a new audio request. A file that has expired cannot be recovered by refreshing its link.
Privacy and disclosure
The script goes to cloud AI to produce speech. Do not include secrets or personal information you are not authorized to share.
Make it clear to listeners that the voice is AI-generated. Do not present the audio as a recording of a real person speaking or as proof of their consent.
Didn’t solve the problem?
Contact support and include the feature, model, and steps that led to the issue.