xAI quietly turned Grok from a chatbot that talks into a full voice platform. The roster went from 5 voices to 28, each one cast for a specific job, and all of them speak 25+ languages natively.
Helix hosts. Every voice on the roster introduces itself in its own voice. No API code is read aloud.
Episode length: 6 minutes 23 seconds.
In this episode:
00:00 Show open
00:15 What actually changed
01:59 The 23 flagship voices
03:13 The five originals, retrained
03:29 How to pick one for support, narration, ads, or education
04:20 Where you can use them: the app, the API, the agent builder
04:55 Speech control tags and voice cloning
05:15 What it costs
05:40 The honest caveats
06:05 Close
Every voice named: carina, zagan, helix, orion, luna, iris, altair, zenith, perseus, helios, lux, kepler, rigel, cosmo, celeste, ursa, sirius, lumen, castor, naksh, atlas, aurora, liora, ara, eve, leo, rex, sal.
Text-to-speech runs $15 per million characters. Realtime speech-to-speech starts at $0.05 per minute.
Note: the consumer Grok app shows a smaller voice list than the developer platform. The full 28 live on the API side.
Read the full post with a playable sample of all 28 voices:
https://blog.cloudhq.net/grok-voices-explained/
From the CloudHQ podcast. More episodes at blog.cloudhq.net. Subscribe if you want helpful, time-saving updates through the week.