Ultravox Realtime offers three ways to customize how your AI agents sound:

​
Built-in Voices

Ready-to-use, high-quality voices for all supported languages. Perfect for getting started quickly.

​
Voice Cloning

Create custom voices from audio samples to match your brand or requirements.
  • Upload a 30-60 second audio sample
  • Get a unique voice that matches your sample’s characteristics
  • Learn more: Voice Cloning Guide

​
Bring Your Own TTS

Integrate your preferred TTS provider for specialized needs or existing investments.
  • Supports ElevenLabs, Cartesia, OpenAI, and many others
  • Complete control over voice generation
  • Learn more: Bring Your Own TTS Guide

​
API Reference

For complete voice management documentation, see the Voices API Reference.