PDF to Markdown

How to use PDF to Markdown

Key features of PDF to Markdown

About PDF to Markdown

PDF to Markdown reads each page with pdf.js and groups text runs into lines by their vertical position. The first line on a page, if under 50 characters, and any line under 35 characters written in capitals become ## headings. Lines beginning with a bullet, a hyphen or an asterisk become list items with the marker normalised to a hyphen. Everything else is written as a paragraph, and the result is offered as a .md download and in an editable box with a copy button.

It suits getting a PDF's text into a README, a wiki, a static site or a notes app that reads Markdown. Bold and italic text, links, images and tables are not detected, and a table comes out as plain lines. The heading rule is simple, so a short first line such as a date may become a heading. A scanned PDF has no text layer and yields only the title and page markers; run it through OCR PDF first.

When to use PDF to Markdown

A vendor's release notes or a changelog arrives as a PDF and you want it in an issue or a pull request that renders Markdown. Convert, copy from the box, and paste; the ## headings and - bullets render as they should, and the page markers show where each PDF page began.

You keep requirements or a checklist in a PDF and want it as a task list. Bullet lines come out as - items, so add [ ] after each hyphen in the box and the list becomes checkboxes.

You want to hand a document to a chat assistant or paste it into a prompt. The output starts with a level-one heading taken from the file name; change it to the document's real title in the box.

When the headings depend on font size rather than position, PDF to Word detects them by size and joins wrapped lines into paragraphs, which this tool does not. If you need a web page rather than Markdown, PDF to HTML uses the same line grouping and writes h2 and p tags. A scan gives only the title and page markers until it has been through OCR PDF.

Tips and troubleshooting

• Failed to convert PDF to Markdown: No password given means the PDF needs a password to open. Remove it first; copy restrictions do not stop the conversion.
• Each PDF line becomes its own paragraph, separated by a blank line. Wrapped lines are not joined, so a five-line paragraph becomes five short ones; join them in the box or use PDF to Word.
• A normal line straight after a bullet is written on the next line with no blank line between, and many renderers then show it as part of the bullet. Insert a blank line after the list.
• Any line starting with a hyphen, an asterisk or a bullet character is treated as a list item, including a negative number such as -5 and a line that begins with a dash. Numbered lists (1., 2.) are not detected and stay as paragraphs.
• Short lines with no lower-case letters count as capitals and become ## headings when under 35 characters, and the first line of every page becomes one when under 50 characters, so a running header repeats as a heading on each page.
• Markdown characters in the text are not escaped: a line beginning with # becomes a heading, paired asterisks or underscores become emphasis, and a line of hyphens becomes a rule. Backslash-escape them in the box if they are meant literally.
• Preview Document loads the .md into a frame; the box under the result is the surer check. Copy Markdown copies the box including edits; Download File saves the original, named after your PDF with an .md extension.
• If the text is going onto a web page and you want markup you can open in a browser, PDF to HTML writes the same text layer as a standalone HTML file.
• If the output holds only a title and page markers, the PDF is a scan. The guide how to make a scanned PDF searchable explains how to add a text layer first.

How PDF to Markdown works in your browser

The file is read into memory and parsed by pdf.js in a worker loaded from cdnjs.cloudflare.com. For each page the script fetches the text runs, rounds their vertical positions and starts a new line when the position moves by more than 6 points. The output string begins with a # heading made from the file name; then, for each page of a multi-page file, a horizontal rule and an italic Page N line; then each line as a ## heading, a - item or a paragraph according to the rules above.

The finished string becomes a Blob of type text/markdown and is passed to showResult for the card, after which the script appends a textarea with the same text and a Copy Markdown button that writes the box's contents with the clipboard API and reads Copied for two seconds. Download File saves the Blob through a temporary object URL. The PDF bytes and the Markdown stay in the tab; the page never sends them anywhere.

Frequently asked questions