PDF to Markdown

Yes — no account, no page limit, and nothing stitched into the output. The extraction runs as code in your browser tab, so a conversion costs us nothing to offer; our paid bank-statement converter is what funds the site.

No. The PDF is opened and read for its text inside this browser tab, and the .md file is written straight to your downloads. What crosses the network is this page's own code, the PDF engine it loads when you press the button, and an anonymous count of how often the tool runs — never the document, and never a line of what it says.

Because a scan is a picture of a page rather than text. This tool reads the text layer a PDF carries — the actual characters and the coordinates they sit at — and an image-only PDF has none, so you get a message saying so instead of an empty file. Run the PDF through OCR first, then convert the result here.

Headings, paragraphs, list items, and code blocks, worked out from type size, position on the page, and whether the font is fixed-pitch. Tables, images, links, and inline bold or italic are not recovered. Neither are list numbers: a numbered list comes back as plain bullets, so numbered clauses or steps have to be renumbered by hand. A title page set entirely in large type can also come through without a heading, because there is no body text on it to measure the heading against.

Lines are read straight across the full width of the page, so two columns interleave: the first line of the left column, then the first line of the right. Table rows flatten into prose for the same reason, and a page saved at an angle can come out in an unexpected order. A simple single-column document comes through close to right — but running heads and page numbers are pulled out along with the body text, and a paragraph broken over a page break arrives as two, so expect a tidy-up either way.