Open notebox

Audio is an easy way to learn on the go, but it fails on technical material like research papers, textbooks, and legal filings. Standard text-to-speech tools read out page numbers, headers, and citation brackets, ruining the flow. They cannot show diagrams or formulas, and there is no way to stop the playback and ask a question when something is unclear. Our project turns dense PDFs into an interactive, visual-first audio experience: - Clean narration: The layout parser strips out running headers, footers, and citation markers while keeping multi-column reading order intact. - Synchronized visual context: High-resolution diagrams, tables, and equations stay visible on screen in sync with the audio, so you can inspect figures without losing your spot. - Interactive podcast call-in: While listening to a generated two-host discussion of your document, you can tap the mic to ask a spoken question. The hosts pause, answer using exact facts from the text, and transition back to the main discussion. - Direct citations and review: In-chat answers link directly to highlighted passages in the PDF. Key points convert into spaced-repetition flashcards with progressive hints for review. - Self-hosted privacy: Users can run the entire stack locally using Meta's open weights, keeping sensitive research and proprietary files private. 🚀 Try the live demo: https://opennotebox.mrxsach.in username : demo password : demo123