Skip to main content
Open Text To Speech → Voice Cloning and choose Add your voice.
  1. Details: give it a name, choose the language the recording is spoken in, and optionally a gender. Confirm that you have the right to clone this voice.
  2. Sample: record with your microphone or upload a file of 5 to 30 seconds.
    • Use one person, speaking clearly, in a quiet room. About 15 seconds is best.
    • Files can be WAV, MP3, M4A or WebM, up to about 2.8 MB. A long WAV is usually larger than that, so use MP3 or M4A.
The voice is ready in under a minute. It plays a short preview on its card, and appears under My Cloned Voices in Text to Speech and Conversation.

​
Who can clone

  • Owners and admins: only they can add or remove a voice, because a cloned voice is a real person’s voice, kept with a record of who agreed to its use.
  • Every member: can listen to the voices and use them.

​
How many

  • Your plan sets the number: 25 on pay as you go, 50 on Pro, 100 on Scale, and custom on Enterprise.
  • Making room: remove a voice to free its place.

​
What we keep

  • The record: the recording, the exact consent text and who agreed are kept with the voice.
  • Removing a voice deletes the voice, its recording and its preview.
These voices are separate from the Voice Clone page for agents’ calls. That page clones on a different engine. A voice made here is for Aero TTS: the console and the API.