Developers on xAI's voice APIs can clone a voice at no extra cost, with speaker verification to stop people cloning others.

Kenning's take

What it does

Custom Voices lets developers clone a voice by recording about a minute of speech in the xAI console, with a model ready in under two minutes. The cloned voice works anywhere xAI's built-in voices do, including the Grok Text to Speech and Voice Agent APIs, with multilingual output. To prevent misuse, the speaker must read a passphrase aloud and xAI checks that the recording matches the same person.

• Clone a voice from about a minute of speech
• Voice model ready in under two minutes
• Works with Text to Speech and Voice Agent APIs
• Multilingual output and speech tags
• Passphrase and speaker-match verification
• REST and WebSocket streaming

Who it suits. Developers building voice agents, narration or branded audio on xAI's APIs.

Key facts

Category

Developer tools

Part of

Grok

Made by

xAI (SpaceXAI LLC)

Released

30 April 2026

How to get it

in the xAI developer console and API

Works on

API, web console

API price

Text to Speech $15 per million characters; speech to speech $0.08 per minute; no extra charge for custom voices

FAQ

Facts from the maker's own pages and the sources listed. Checked 2026-10-11.