When you use LiveAvatars in FULL Mode, we deal with the common setup pains with setting up realtime chat, including:

• Voice Activity Detection (VAD)
• Speech-to-Text (STT) / Transcription
• Large Language Models (LLM)
• Text-to-Speech (TTS)

FULL mode is recommended for developers that want to:

1. Delegate the orchestration/infrastructure of managing the WebRTC structure.
2. Avoid building and maintaining a complex real-time AI pipeline across audio input, inference, and audio output.
3. Ship faster without worrying about model coordination, streaming latency, or state management.

By delegating the full real-time stack to LiveAvatars, you’re free to focus on what matters most: designing the best possible product and user experience for your specific use case.

Getting started

Follow our quick start guide, where we walk through how to create a session token in FULL mode, and start the session. We recommend checking out our architectural guide for how we recommend setting up FULL mode.