🚨 Someone built a free, open-source NotebookLM that runs entirely on your machine. No Google account. No subscription. No data leaving your computer. Same concept. Better privacy. Zero cost. It's called SurfSense. 4,800 GitHub stars. And it does everything NotebookLM does — plus things NotebookLM can't. Here's what NotebookLM actually is for anyone who missed it. Google's NotebookLM lets you upload documents — PDFs, notes, articles, research papers — and chat with them. Ask questions. Get summaries. Have it explain complex sections. It even generates podcasts from your documents with two AI hosts discussing the content. It went viral last year. Millions of users. Because the use case is genuinely transformative for research and learning. The problem: your documents go to Google. Your research. Your notes. Your private materials. All processed on Google's infrastructure. Subject to Google's data policies. Connected to your Google account. SurfSense runs locally. Your documents stay on your machine. Here's what it can actually do: → Chat with any document — PDFs, research papers, articles, notes, web pages → YouTube integration — paste any YouTube URL, it transcribes and indexes the video → Web page capture — browser extension captures any page you're reading for later querying → Podcast generation — same two-host AI podcast feature that made NotebookLM famous. Free. Local → Multi-document reasoning — ask questions that span multiple documents simultaneously → Citation tracking — every answer traces back to the specific source document and page → Any LLM — OpenAI, Claude, Gemini, or any Ollama local model. Full offline capability with local models Here's what it does that NotebookLM can't. NotebookLM only accepts documents you explicitly upload. SurfSense captures everything — the article you're reading right now via browser extension, the YouTube video you just watched, any webpage you visit. Your entire research trail becomes searchable, not just the files you remember to upload. And the local model support is the differentiator. With Ollama, your documents never touch any external server — not Google's, not OpenAI's, not Anthropic's. Processing happens entirely on your hardware. Here's the wildest part. The podcast generation is genuinely impressive. Two AI hosts with distinct personalities discuss your documents — one curious and questioning, one more analytical and explanatory. 10-minute deep dives on any paper or article you upload. The same feature that made NotebookLM go viral. Free. Unlimited. Running locally. Here's what one install gives you. A personal knowledge base that captures everything you read and watch. Fully searchable. Chattable. Podcast-able. With no external data sharing whatsoever. One command to start. Open localhost:3000. Upload your first document. Start chatting. 4.8K GitHub stars. 511 forks. 1,105 commits. MIT License. 100% Open Source. GitHub link in the comments 👇