Grok Launches Voice Cloning: Record in One Minute to Create Your Own AI Voice

According to Dynamix Beating monitoring, xAI has launched Grok Custom Voices and Voice Library. Users can record a voice sample in the xAI console, generate their own voice_id, and then integrate with Grok TTS or Voice Agent API for use cases such as customer service agents, content creation, game characters, audiobook narration, and more.

This feature is not a simple audio upload for cloning. Users need to provide voice verification by reading a short sentence. The system will use STT for real-time transcription, compare the voice characteristics of the verification recording and the full recording, and only generate the voice if the speaker is confirmed to be the same person. xAI claims that this process prevents the cloning of someone else's voice using pre-existing recordings.

Currently, Custom Voices are only available in the United States, except for Illinois. The console allows a maximum of 30 free custom voice creations, and API creation capability is only available to Enterprise teams. The custom voices themselves do not incur additional charges, but usage of the voice API is billed based on consumption: Realtime is priced at $3.00 per hour, and Text to Speech is priced at $4.20 per million characters.