network-exposure. The README notes the UI (8502) and API (5055) are published on all interfaces with no password by default; do not run it on a shared network as is.
Who it is forA researcher whose documents are too sensitive to upload to Google's cloud.
The problemNotebookLM only runs on Google's servers with Google's models.
The outcomeOpen Notebook running at localhost:8502 with your own encryption key.
What you need first
- Docker Desktop (macOS, Windows) or Docker Engine with the Compose plugin (Linux)
- An API key from an AI provider, or a local model server such as Ollama
- At least one language model and one embedding model configured in the UI
Steps
- Run
mkdir open-notebook && cd open-notebookand download docker-compose.yml with the README's curl command - Change OPEN_NOTEBOOK_ENCRYPTION_KEY to a long random secret, for example from
openssl rand -hex 32 - Start it with
docker compose up -d - Wait about 30 seconds and open http://localhost:8502
Prompt for your agent
Set up lfnovo/open-notebook on this machine with the official docker-compose.yml, generate an encryption key, start it and confirm the UI loads on port 8502. Warn me if the ports are exposed beyond localhost.
Caveat. This card is written from the lfnovo/open-notebook README, not from a test run by Deduped. Before you start, make sure you have Docker Desktop or Docker Engine with the Compose plugin. Check the repo for newer instructions before you rely on these steps.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for lfnovo/open-notebook and "Self-host a private NotebookLM alternative with Docker"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer