Thanks — I’ll explain LiveKit avatars. LiveKit doesn’t provide a native avatar generator. Instead, LiveKit integrates visual avatars by connecting Agents to third‑party avatar providers via plugins. That lets an Agent drive a synchronized video avatar (lip sync + expressions) using the Agent’s TTS/audio output and publish the avatar as a video track into a LiveKit room. Key points • Avatars are provided by third‑party services (examples: Anam, Avatario, D-ID, LiveAvatar/HeyGen, TruGen, bitHuman). • Integration is via provider plugins that manage an AvatarSession which streams video frames and syncs with TTS. • Avatar video tracks appear in the LiveKit room like any other video track and can be shown in starter apps (for example with @livekit/components-react). • Setup requires installing the provider plugin, setting the provider API key(s) (typically in

.env

), initializing an

AvatarSession

, and starting it with the AgentSession/room. • If an avatar service fails, the agent continues as audio‑only (graceful degradation). If you want, I can: • Walk through setup for a specific provider (which one are you using or considering?), or • Share example code for initializing an

AvatarSession

for a chosen provider. Sources: Anam virtual avatar integration guide | LiveKit Documentation | Avatario Realtime Avatar integration guide | LiveKit Documentation | LiveAvatar virtual avatar integration guide | LiveKit Documentation | TruGen.AI realtime avatar integration guide | LiveKit Documentation | bitHuman virtual avatar integration guide | LiveKit Documentation, with interactive elements