Convert to Markdown for notes and LLMs. Everything runs locally in your browser — your files are never uploaded to a server, no registration is required, and the tool is completely free.
How to PDF to Markdown
Upload the PDF you want as Markdown.
Run the conversion; headings and lists are inferred from the layout.
Review the Markdown shown on screen.
Copy it straight into your notes, docs or an LLM prompt.
Features
Infers Structure From Layout: A PDF has no headings — only glyphs at a position and a size. The most common text size on each page is taken as body copy, and anything materially larger becomes a heading, with the level scaled by how much larger it is.
Detects Lists and Paragraphs: Lines beginning with a bullet glyph or a number become list items, runs on the same baseline are joined into one line, and a vertical gap wider than a line becomes a paragraph break.
Clean Output for LLMs: Markdown is far cheaper and clearer to feed a language model than raw PDF text, and it keeps the heading structure that tells the model how the document is organised.
Runs Entirely In Your Browser: The document is parsed locally. Nothing is uploaded — which matters if the reason you want Markdown is to paste it somewhere else.
Frequently asked questions
How accurate is it?
Very good on linear, text-led documents — reports, papers, manuals, contracts. It is a layout heuristic, not a full document-structure model, so heavily designed pages with multiple columns, sidebars or complex tables will need tidying by hand afterwards.
Why are my tables a mess?
Tables have no structural markup in most PDFs; they are text positioned in a grid. Recovering them reliably needs a different approach — try the PDF to Excel tool, which is built for tabular data.
Will it read a scanned document?
No. A scan is an image with no text to extract. Run it through OCR PDF first to add a text layer, then convert.
Why did a heading come out as body text?
The document probably signals that heading with weight or colour rather than size. Size is the only reliable cross-document signal, so a bold heading set at body size will not be detected.