interview-agent

Voice interview practice agent built with LiveKit Agents. The agent’s name is Rachel and it runs a structured interview:

• Intro: collects your name, current status/background (1 sentence), and the job title you’re interviewing for
• Past experience: collects exactly 2 concrete examples with up to 3 follow-ups per example
• Auto-end: ends if you’re silent for too long (and after you ask to stop)

What to run

main.py is the entrypoint. After setup, run:

python3 main.py dev

If you’re unsure what commands are available, once dependencies are installed you can also run:

python3 main.py --help

Setup

Create a virtual environment and install dependencies:

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Configuration (.env)

This project uses python-dotenv, so you can create a .env file in the repo root.

At minimum, you need an OpenAI key:

OPENAI_API_KEY=your_openai_api_key

To connect to a LiveKit server (Cloud or self-hosted), you typically also need:

LIVEKIT_URL=wss://your-livekit-host
LIVEKIT_API_KEY=your_livekit_api_key
LIVEKIT_API_SECRET=your_livekit_api_secret

Notes / troubleshooting

• ModuleNotFoundError on startup: you didn’t install dependencies yet; run the setup commands above.
• Auth errors: confirm OPENAI_API_KEY (and LiveKit env vars if you’re using a LiveKit server).
• Audio issues: make sure your microphone permissions are enabled in whatever client you use to join the LiveKit room.