Thanks — I can help with that. You’re asking how to integrate a LiveKit avatar into your frontend; I’ll walk you through the general approach and can provide code specific to your stack once you tell me which frontend framework and avatar provider you’re using. Quick summary of how avatars work (what you need to render) • The avatar is published by a separate participant called the avatar worker (participant kind = Agent) that joins the room when an AvatarSession starts. • The avatar worker has an attribute

lk.publish_on_behalf

set to the agent’s identity — use this to identify the avatar worker. • Render the avatar’s audio and video tracks (from the avatar worker), not the agent participant’s tracks. If you want the short integration checklist 1. Identify the avatar worker in the room: ◦ Find the agent participant (agent identity with no

lk.publish_on_behalf

), then find the participant where

lk.publish_on_behalf === agent.identity

. 1. Get the avatar worker’s audio and video tracks. 2. Attach the video track to a DOM/video component and the audio track to an audio element (or let the SDK handle playback). 3. Optionally: use the provided React helper

useVoiceAssistant

which already finds the avatar worker and returns

agent

,

audioTrack

, and

videoTrack

. What I need from you to give exact code • Frontend framework / platform (choose one): React (web), Vue, Angular, React Native, Flutter, SwiftUI, Android, or plain JS. • Are you using any LiveKit helper components or the plain LiveKit client SDK? • Which avatar provider/plugin are you using (examples: Anam, Tavus, Hedra)? Do you already have the AvatarSession set up on the backend? If you want a React example right away • I can provide a minimal React snippet showing how to: a. Find the avatar worker via participant attributes. b. Attach the

VideoTrack

to a <video> element (or LiveKit React components). Tell me if you prefer that and which UI approach you use (plain DOM elements vs. LiveKit React components). If you want links or starter apps • I can point you to example starter apps (Next.js, React Native, Flutter, SwiftUI, Android) that already handle avatar rendering. Which option would you like me to show next? (Pick one) • A minimal plain-React + LiveKit client snippet • A

useVoiceAssistant

-based React example • A Vue/Angular/plain-JS example • Mobile example (React Native / Flutter / native) Sources: Virtual avatar models overview | LiveKit Documentation | Anam virtual avatar integration guide | LiveKit Documentation