AI Note Taker for LiveKit

An installable Python package that runs a LiveKit agent which listens to room audio, transcribes with Deepgram and AWS transcribe, and publishes concise meeting notes/summaries using Groq, OpenRouter(ollama), or OpenAI-compatible models.

Installation

pip install .
# or once published:
pip install livekit-ai-note-taker

Configuration

Set the following environment variables (a .env file is supported via python-dotenv):

• LIVEKIT_URL – LiveKit WebSocket URL (e.g. wss://example.livekit.cloud)
• LIVEKIT_API_KEY / LIVEKIT_API_SECRET – LiveKit credentials
• DEEPGRAM_API_KEY – Deepgram STT key
• PROVIDER – groq, openrouter, or openai
• GROQ_API_KEY / GROQ_MODEL_NAME – when PROVIDER=groq
• OPENROUTER_API_KEY / OPENROUTER_MODEL_NAME / OPENROUTER_BASE_URL – when PROVIDER=openrouter
• OPENAI_API_KEY – when PROVIDER=openai
• SUMMARY_INTERVAL – seconds between rolling summaries (default 150)
• PROMPT_TYPE – small or big prompt template (default small)
• AGENT_NAME – LiveKit agent registration name (default meeting_summarizer)
• BACKEND_URL – webhook endpoint that receives the final meeting summary (default http://localhost:8000/meetings/meeting-note-webhook)
• LIVEKIT_SUMMARIZER_DISABLE_SSL_VERIFY – set to true to ignore TLS verification for aiohttp (development only)

Running the worker

livekit-ai-note-taker

The command validates configuration, connects to LiveKit, and starts publishing summaries to the room via publish_data on the local participant.

Using as a library

import asyncio
from livekit_ai_note_taker import run_worker

asyncio.run(run_worker())

You can also import NoteTakerAgent if you want to wire custom events or prompts in your own application. Logging is left to the host application; configure it as needed (e.g. logging.basicConfig(level=logging.INFO)).