anam

Anam healthcare intake assistant

A LiveKit Agents demo that guides patients through a medical intake form using voice and an Anam lip-synced avatar. The agent (Liv) collects information one field at a time and updates the form in real time via RPC to the frontend. The demo includes both a Python and a TypeScript backend; run whichever you prefer. The frontend works with either.

Features

• Voice intake flow: Liv asks for each form field in order (name, DOB, address, phone, emergency contact, medications, allergies, reason for visit)
• One question at a time: Confirms each answer before moving to the next field
• Real-time form updates: Agent updates the on-screen form as the user speaks via RPC
• Anam avatar: Lip-synced video avatar powered by the Anam plugin
• Voice pipeline: Deepgram STT, OpenAI LLM, and ElevenLabs TTS (via LiveKit Inference)
• Form submission: Final confirmation loop, then submit from the frontend

Prerequisites

• For the Python backend: Python 3.10+, uv
• For the TypeScript backend: Node.js, pnpm
• For the frontend: Node.js, pnpm
• LiveKit account
• Anam API key

Installation

1. Clone this repository and go to the anam demo

   git clone <repository-url>
   cd complex-agents/avatars/anam

2. Install the backend you want to use (or both).

   • Python backend

     cd agent-py
     uv sync
     cd ..

   • TypeScript backend

     cd agent-ts
     pnpm install
     cd ..

3. Install frontend dependencies

   cd frontend
   pnpm install
   cd ..

4. Set up environment for your chosen agent

   • Python: In agent-py, copy .env.example to .env.local and add your LiveKit and Anam credentials.
   • TypeScript: In agent-ts, copy .env.example to .env.local and add your LiveKit and Anam credentials.

5. Set up environment for the frontend
   In frontend, copy .env.example to .env.local. Add the same LiveKit credentials and set AGENT_NAME=Anam-Demo so the frontend dispatches to the agent.

Configuration

Enter your environment variables in the .env.local file:

LIVEKIT_URL=wss://your-project.livekit.cloud
LIVEKIT_API_KEY=your_livekit_api_key
LIVEKIT_API_SECRET=your_livekit_api_secret
ANAM_API_KEY=your_anam_api_key

Get your Anam API key from the Anam dashboard. You can get your LiveKit environment variables from the LiveKit dashboard or load them from the LiveKit CLI:

lk cloud auth
lk app env -w -d .env.local

Run that from either agent-py or agent-ts depending on which backend you use.

Frontend (frontend/.env.local)

LIVEKIT_API_KEY=your_livekit_api_key
LIVEKIT_API_SECRET=your_livekit_api_secret
LIVEKIT_URL=wss://your-project.livekit.cloud
AGENT_NAME=Anam-Demo

Use the same LiveKit values as the agent. AGENT_NAME=Anam-Demo matches both backends so the frontend connects to whichever agent you run.

Usage

Run one backend (Python or TypeScript) and the frontend. Do not run both backends at the same time.

Python backend

1. One-time: download Silero VAD and the turn detector model

   cd agent-py
   uv run python src/agent.py download-files

2. Start the agent

   uv run python src/agent.py dev

The agent registers with your LiveKit project as Anam-Demo.

TypeScript backend

1. One-time: download Silero VAD and the turn detector model

   cd agent-ts
   pnpm run download-files

2. Start the agent

   pnpm run dev

The agent registers with your LiveKit project as Anam-Demo.

Frontend

1. Open a new terminal and go to the frontend:

   cd frontend

2. Start the dev server:

   pnpm dev

3. In your browser, open:

   http://localhost:3000
   

4. Start a session and complete the intake form with Liv.

Project structure

anam/
├── agent-py/              # Python backend (optional)
│   ├── src/agent.py       # Intake agent and Anam avatar session
│   ├── pyproject.toml
│   └── .env.local
├── agent-ts/              # TypeScript backend (optional)
│   ├── src/               # Agent entry and intake logic
│   ├── package.json
│   └── .env.local
└── frontend/
    ├── app/               # Next.js app router
    ├── components/app/    # Session, avatar, intake form UI
    └── .env.local         # LiveKit credentials + AGENT_NAME

Both agents use the same agent name (Anam-Demo) and the same function tools (update_field, get_form_state, submit_form) to call RPCs on the frontend so the form stays in sync with the conversation.

Built with

• LiveKit Agents - Agent framework
• Anam plugin - Lip-synced avatar
• LiveKit Inference - STT, LLM, TTS models