Livekit Agent

Tavus offers integration with the Livekit Agents framework, an open-source framework for building conversational agents by Livekit. You can easily add Tavus Replicas to your Livekit agents and give them a video layer.

You can keep your Livekit workflow as-is and just create a new Tavus conversation with certain settings.

​

Step 1: Setup Tavus Replica

First, you need to find a Tavus replica to use with your Livekit Agent. This can be any replica that’s stock, or your own replica.

​

Step 2: Create a Tavus Persona

After finding a Tavus replica, you need to create a persona with certain settings. Specifically, you need to set the pipeline_mode to echo, and define a transport layer under layers, setting the transport_type inside to livekit.

​

Step 3: Initiate the Conversation

Once you got back a persona ID and have your replica ID ready, create a conversation. You need to add the following parameters to your conversation properties:

"livekit_ws_url": "your-livekit-websocket-url",
"livekit_room_token": "your-livekit-room-token"

​

Step 4: Begin the Conversation

After you POST a conversation with the above settings and the correct persona, the replica will join the Livekit room and begin listening to participants that are of kind rtc.ParticipantKind.PARTICIPANT_KIND_AGENT to pipe the audio output received by them with the video layer attached and synced.