avatar

Render a lip-synced talking-head video from audio using an avatar provider.

Provider

Tavus — provider: tavus

Product links: tavus.io · Tavus API

Example

{
  "name": "host_avatar",
  "type": "avatar",
  "inputs": ["mix_audio"],
  "config": {
    "provider": "tavus",
    "interrupt": true,
    "config": {
      "apiKey": "<tavus-api-key>",
      "replicaId": "<replica-id>",
      "personaId": "<persona-id>"
    }
  }
}

Config

Field	Default	Description
provider	—	tavus
interrupt	true	Stop avatar speech when participant talks
config.apiKey	—	Tavus API key
config.replicaId	—	Replica (face/voice) ID
config.personaId	—	Echo-mode persona with microphone enabled
config.participantName	Tavus-avatar-agent	Name to subscribe to in the Daily room

Outputs audio + video — wire into video_mixer or sinks.

Pricing

$0.05/min while the avatar session is active. Node pricing.