About
I built a PDF to Markdown converter that runs 100% client-side. There's no upload, no sign-up, and no limits — the PDF is parsed in JavaScript on your device, and nothing is sent to a server. You can actually disconnect from the internet after the page loads and it still works, which makes it usable for confidential documents.
The conversion reads text, font sizes, and layout to reconstruct headings, lists, and paragraphs as Markdown, rather than just dumping a wall of text. Output is copy-to-clipboard or a downloadable .md file, ready for Obsidian, Notion, GitHub, Hugo, or feeding into an LLM.
The obvious tradeoffs: complex tables get flattened to text, images aren't extracted, and scanned PDFs (no text layer) come out empty since there's no OCR step. If you need higher-fidelity ML conversion, tools like Marker, MinerU, or Docling are better — but those need a local install and often a GPU. I wanted something with zero setup for the quick everyday case.
There are also sibling tools on the site (Markdown to PDF, DOCX to Markdown, PDF to JSON, etc.), all browser-only.