Open Notebook is a tool where you can upload your own sources - PDFs (textbooks, research papers), Github repos, documentation websites, and other URLs. Once uploaded, you can either chat with an AI about them or run “apps” on top of those sources.

Apps can be anything. Right now, I’ve built the following: - turn articles into short video overviews - create a mind map of the source - create summaries or other reports about the source - generate quizzes or flashcards to help study the source

You can also remix existing apps or create one from scratch.

I built this because I wanted to make self-learning less painful. I enjoy diving into new subjects (electronics, space, math), but the usual options feel limited: long lectures, repetitive ChatGPT chats, or slogging through PDFs and textbooks.

NotebookLM looked promising in this space, but I found it too rigid - the outputs weren’t bad, but they weren’t adaptable to how I wanted to learn.

Open Notebook takes a different approach: it exposes the full workflows behind the AI outputs (LLM prompts, steps, formats). That means you can see exactly how a video, podcast, or summary is generated, change the process, or share your own workflows with others.

That is why I consider Open Notebook to be the "open source" version of NotebookLM.

I’d be interested in feedback on what’s useful, what’s missing, or how you’d want to adapt it for your own learning.