Open Notebook
The open-source, privacy-first research notebook — bring your own AI models, feed in documents and media, then chat and generate podcasts, all self-hosted. A NotebookLM alternative you own.
NotebookLMAbout Open Notebook
Open Notebook is what you reach for when you love what Google's NotebookLM does but not the fact that your research sits on Google's servers. It's a self-hostable research notebook that turns documents, notes, videos and web pages into a source you can chat with, summarize and reason over.
The key difference is control: bring your own AI models — OpenAI, Anthropic, Ollama and others — so you decide which provider (or local model) sees your data, and how much you spend. Nothing is locked to a single vendor.
It's MIT-licensed and built to run on your own machine or server with Docker, and it can even generate podcast-style audio overviews the way NotebookLM does. For researchers and teams who want that workflow without giving up their data, Open Notebook is the leading open answer.
Key features
- Bring your own model — Use OpenAI, Anthropic, Ollama or other providers — you choose which model sees your notes and data.
- Any source — Feed in PDFs, notes, links, YouTube videos and web pages, then chat and search across all of them.
- Podcast generation — Turn your sources into a podcast-style audio overview, the way NotebookLM does.
- Self-hosted & private — Run it locally or on your own server with Docker — your research never leaves your control.
The honest read
- Privacy-first — your documents stay on your own infrastructure
- Bring your own AI provider or run a local model
- MIT licensed and fully open source
- Supports many source types, including video and web pages
- Generates podcast-style audio overviews
- Requires you to supply your own AI API keys or local model
- Self-hosting takes more setup than a hosted product
- UI is less polished than Google's NotebookLM
- Audio/podcast quality depends on the models you plug in
FAQ
Yes — Open Notebook is free and open source under the MIT license. You only pay for whichever AI provider you connect (or nothing, if you run a local model).
Open Notebook is self-hostable and privacy-first — your data stays on your own infrastructure and you choose which AI model to use, while NotebookLM is a closed, Google-hosted product.
You can bring your own provider, including OpenAI, Anthropic and local models via Ollama, so you control cost and which model sees your data.
Yes — it can produce podcast-style audio overviews from your sources, using the models you connect.
Yes — it's built to run locally or on your own server with Docker.
Python
TypeScript
SurrealDB
DockerSimilar open-source projects
Open-source, local-first workspace for docs, wikis and projects.
End-to-end encrypted, local-first knowledge base built around objects.
Open-source embedding database for AI apps and RAG.