Speech settings allow you to control how your AI agent sounds during conversations.
You can choose from multiple voice providers, customize models, and even add background sounds for more natural phone interactions.

​
Voice Selection

You can select voices from different speech providers directly within the platform.
Supported providers include:
  • OpenAI
  • Cartesia
  • Rime AI
  • ElevenLabs
You can also:
  • Browse available voices from each provider in the dropdown.
  • Test a voice before assigning it to your agent.
  • Add custom voices (e.g., cloned or uploaded voices in supported providers).
Agent Speech Settings
Selecting the right voice helps make your agent sound more natural and aligned with your brand.

​
Model Selection

Each voice provider offers multiple speech synthesis models (for quality, latency, or cost optimization). From the Model dropdown, you can:
  • Select the model best suited for real-time phone calls.
  • Balance between high-quality natural speech and low-latency response times.
Agent Speech Settings
For production use-cases like live calls, we recommend models optimized for low latency.

​
Background Sounds

To make calls feel more authentic, you can add optional background ambience. Available options include:
  • Office Ambience
  • Keyboard Typing
Agent Speech Settings
You can enable or disable background sounds per agent, depending on your use case.

​
Summary

  • Choose voices from providers like OpenAI, Cartesia, RhymeAI, ElevenLabs.
  • Add custom voices if supported by the provider.
  • Select the appropriate speech model for performance and quality.
  • Optionally enable background sounds (office, typing, neutral) to make calls sound more realistic.
Speech settings help you design the tone, personality, and realism of your AI agent.