LiveKit's integration with Tavus is focused exclusively on enabling hyper-realistic interactive avatars for conversational video AI agents. This integration supports audio-to-video synthesis for avatar generation and real-time avatar rendering, but does not include perception models or perception capabilities.
There is no support for perception models within the LiveKit-Tavus integration; the functionality is limited to avatar generation and video synthesis.
Sources:
Tavus virtual avatar integration guide | LiveKit Docs |
Bringing AI avatars to voice agents