Hey! 👋 We're Carlos, Ale, and Naty from Colombia, and we built TokDown because we were tired of "AI PDF converters" that do one thing: push everything through an LLM. Here's the thing: most PDFs already have a text layer. Sending them to an LLM is slow, costs money, and risks the model hallucinating words that aren't even in your document. That felt wrong to us. So TokDown takes the honest approach: it checks first whether your PDF already has text. If it does, it extracts it character-for-character, instantly, for free, with zero hallucinations. Only true scans go through OCR, and we always tell you which path your file took. Almost no one does that. A few decisions we're proud of: +Zero data retention: Your file is processed and discarded immediately. Scans processed via third-party OCR are discarded right after. No signup, no accounts. +Deterministic-first: Text is text—no AI guessing when it doesn't need to. +Full transparency: Every result clearly labels whether it was "Deterministic" or "AI (vision)". +Honest limits: 5 pages per file for now, while we keep our OCR bill sane. +Trilingual: The web app works seamlessly in Spanish, English, and Portuguese. +Chrome extension: Convert without ever leaving your browser. It's free and we'd love your honest feedback. What kind of PDFs give you the biggest headache to convert? Drop a comment below—we'll read every single one!

TokDown

PDF → Markdown, instantly. No login, no storage.

PDF → Markdown, instantly. No login, no storage.