Overview

Avaluma avatars give voice agents a face. Only the agent's voice output is required to control the avatar.

The avatars are compatible with all common LLM and TTS providers. This makes the avatars versatile. For example, classic voice agent pipelines consisting of speech-to-text, large language models, and text-to-speech can be connected, as well as modern real-time LLMs with voice input and output.

For easy connection to existing systems and low latency through WebRTC, avaluma avatar uses LiveKit for integration and communication (more integrations to follow).