Give your voice agent a face.
Keep your voice pipeline. Add a bitHuman avatar, then test a conversation in your app.
Start with your framework
Each guide explains the setup and shows how the avatar connects.
LiveKit
Add the avatar to your existing agent and room.
- Where the avatar renders
- Your Python worker or bitHuman cloud
Choose the local model or cloud avatar path in the guide.
Pipecat
Render the speech from your voice pipeline as an avatar.
- Where the avatar renders
- Your Python process
The bitHuman community integration goes after your speech service.
Your own voice pipeline
Send reply audio to a cloud avatar in your LiveKit room.
- Where the avatar renders
- bitHuman cloud
Use the REST session path when your bot is not a LiveKit Agents worker.
You need a bitHuman account and API secret. Your voice and transport services may need their own accounts. Each guide lists what to install.
Check a full conversation
- 1. Get the first reply. Use a sample avatar and your existing voice configuration. Check that you can see the avatar and hear one audio output.
- 2. Try your app's behavior. Speak, interrupt a reply, reconnect and end the session. Check the avatar follows your conversation and stops when the session ends.
- 3. Check your deployment. Measure the complete pipeline on your intended device or server. Include your voice, transport and avatar usage when estimating cost.
For a cloud avatar, keep your account secret on the server and give the client only session credentials. Follow the guide's token setup.