A PDF stores glyphs at coordinates, not headings and paragraphs. This tool reads that text layer in your browser and rebuilds the structure it can infer: lines drawn in a larger font become headings, lines that start with a bullet or a number become list items, runs of ordinary lines become paragraphs, and clickable link areas become Markdown links. The result is plain text you can paste into a notes app, a wiki, a chat prompt or a retrieval pipeline.
Nothing is uploaded. The file is opened with a PDF reader that runs inside the page, and the Markdown is produced on your device. A contract, a research paper or an internal report stays where it is.
How headings and lists are decided
The tool measures the font size of every line and finds the size most of the document is set in. That is the body size. A line at roughly one and three quarter times the body size or larger becomes a level one heading, one at about one and a third times becomes level two, and a line just a little larger than the body becomes level three. A short heading that wraps onto a second line at the same size is joined back into one.
A line that opens with a bullet character, a dash, or a number or single letter followed by a full stop or bracket is treated as a list item. Numbered markers keep their numbers; everything else becomes a dash. A following line that sits further from the margin than the marker is folded into the item as continuation text.
Paragraphs, page breaks and links
Body lines are joined with single spaces until something breaks the flow: a larger vertical gap than the usual line spacing, a first line indented past the one above it, or a heading or list. A word hyphenated across a line end is stitched back together when the next line starts in lower case.
Each page is emitted in order. By default a horizontal rule separates pages so you can see where one ended; turn that off in Options if you want continuous text. Where the PDF carries a clickable link over a piece of text, that text is written as a Markdown link with the address the PDF points to. Links that exist only as visible blue text, with no link area behind them, come through as plain text.
What does not survive the trip
Tables have no structure in a PDF text layer. The cells arrive as individual lines in whatever order the file drew them, so a table comes out as a run of short lines, not as a Markdown table. Multi-column layouts are read from the top of the page down by baseline, which can interleave the left and right columns; single-column documents convert cleanly.
Bold and italic are not recorded, because the text layer exposes a font name, not a weight, and guessing from names is unreliable. Images, charts and footnote markers are dropped. Headers and footers that repeat on every page are kept, since the tool cannot tell them from body text.
Scanned documents
A scanned PDF is a set of page images with no text layer, so there is nothing for this tool to read. It reports that no text was found instead of producing an empty file. Run such a document through an OCR program first, then convert the searchable PDF it produces. A PDF where only some pages are scans converts the pages that do have text and tells you how many were skipped.
Frequently asked questions
Is my PDF uploaded anywhere?
No. The file is opened and read inside your browser, and the Markdown file you download is assembled on your own device. Nothing is sent to a server.
Why did some headings come out as ordinary text?
Headings are recognised by font size alone. A heading set in bold at the same size as the body, or only a point or two larger, is not distinguishable from body text in a PDF, so it stays a paragraph. Add the hash marks by hand in those cases.
Why is my table a jumble of short lines?
A PDF does not store rows and columns, only text at positions. The tool reads each line in turn, so cells become lines. Rebuild the table by hand or copy the values into a spreadsheet, which is usually quicker than fixing the text.
Can it read a scanned PDF?
No. A scan has no text layer. The tool tells you when it finds none. Use an OCR program to add a text layer first, then convert the result here.
Two columns of text are mixed together. Why?
Lines are ordered by their vertical position on the page, so a line from the left column is followed by the line beside it in the right column. Documents with a single column do not have this problem.
Where do the links come from?
From link annotations in the PDF, the invisible clickable rectangles a word processor adds when it exports. Text that merely looks like a link, with no annotation behind it, stays as plain text, including a bare URL.
What is the horizontal rule between pages for?
It marks where each PDF page ended, which is useful when you need to cite a page number or check the conversion against the original. Switch it off in Options to get continuous text with a paragraph break instead.
Related Utilities You Might Need
Scanned PDF Converter
Simulate real scanned sheets from loaded PDF pages by adding skews, blurs, and scanner noise.
Merge PDF
Combine several PDFs into one file in any order. Runs entirely in your browser, so nothing you merge is ever uploaded.
Split PDF
Cut a PDF into separate files, every few pages or by an exact range. Processed on your device, never uploaded.