Hi HN. This is a personal side project I've been building for about six months on top of the open-source Open Notebook platform (https://github.com/lfnovo/open-notebook).

Notebooker saves the stuff you'd otherwise lose in tabs and bookmarks - links, PDFs, audio, video - and makes it useful later: chat with a notebook and get answers that cite the exact source, turn a reading list into a podcast episode (private RSS feed, works in any player), or generate study material from your own sources.

The part I've had the most fun with is a plugin engine for creation types - flashcards, charts, infographics, mindmaps, textbooks, essays, slideshows, timelines, wikis are each plugins, and new ones can be added without touching the core. Everything I've built on top of open-notebook has either been submitted upstream as PRs or is open source at https://github.com/Notebooker-ai.

Decisions this crowd might care about:

- Bring your own AI keys (OpenAI, Anthropic, local/OpenAI-compatible endpoints) or use the built-in defaults. I enjoy testing Cloudflare Worker AI models. Nothing you save is used for training. You can deploy you own OpenAI compatible endpoint to play with at https://open.notebooker.ai

- Bring your own S3-compatible storage (R2, Spaces, AWS) if you want your files in a bucket you control.

- Webhooks in and out — anything that can POST JSON can trigger a workflow, and workflows can POST anywhere. Notebook also processes incoming RSS feeds and generates its own

- Export everything or delete your account with a click. There's a view-only demo notebook at https://app.notebooker.ai/demo.

Happy to answer any questions. I'm expecting some hiccups in releasing, feel free to report bugs or feedback.