NotebookLM without the cloud lock-in or the containers

What it does

KnowNote is an Electron desktop app that ingests PDFs, Word docs, PowerPoints, and web pages into a local SQLite database backed by sqlite-vec for semantic search. It then runs RAG-powered Q&A against your documents through whichever LLM provider you prefer—OpenAI, DeepSeek, Ollama, or others—surfacing the exact sources it used. The app also generates structured notes and mind maps from your material, keeping everything stored locally on Windows or macOS.

The interesting bit

The author built this as a first open-source project after finding that nearly every NotebookLM alternative on GitHub demanded Docker, which is a deal-breaker for non-backend users. By stuffing SQLite, vector search, and provider-agnostic LLM calls into an Electron shell, KnowNote trades scalability for immediacy: you download it, add files, and start asking questions without touching a terminal.

Key highlights

• Local-first by design: all documents and vectors live in SQLite on your machine, with optional offline use when paired with a local LLM like Ollama.
• Pluggable LLM providers: abstracts OpenAI, DeepSeek, Ollama, and others behind a unified interface.
• Source traceability: RAG answers include precise references back to the original document chunks.
• Broad file support: parses PDF, .docx, .pptx, and web pages automatically.
• Mind maps and structured notes: one-click generation from imported content.

Caveats

• Early-stage and explicitly rough around the edges; this is the author’s first open-source release.
• Linux is not mentioned as a supported platform; only Windows and macOS builds are listed.
• Several headline features—audio transcription, quiz generation, and one-click PPT export—are still in development or planned.

Verdict

Worth a look if you want a private, desktop-bound alternative to NotebookLM and would rather not containerize your note-taking workflow. Skip it if you need a battle-tested, cross-platform server deployment or deep collaborative features.