PDF to Markdown Converter (.md)
Turn the text of a PDF into a clean Markdown file with # headings, - lists and bold emphasis
Your files stay with you. Conversion happens inside the browser; no file is ever uploaded to a server.
Did this tool do the job?
Thanks, your feedback came through.
How it works
Choose your PDF in the box (tap or click) and press "Convert to Markdown"; you receive a .md file named after the PDF. The converter pulls the words out of the PDF's text layer, stitches broken lines into paragraphs and marks the structure in Markdown: heading sizes become #, ## and ### from largest to smallest, lines starting with "•", "–" or "-" become "- " list items, lines starting with "1." become a numbered list, and a paragraph set entirely in bold is wrapped in **double asterisks**. A concrete case: you want an AI chat to summarize a 40-page product manual, but when you paste the PDF the headings vanish and page numbers land in the middle of sentences. Convert it here: repeated lines such as "User Guide – Page 12" are dropped, sections are separated by # headings, and the model sees how the document is organized. The same file opens directly in Obsidian, GitHub, VS Code or Notion's import. The limit: tables become plain lines rather than Markdown tables, and images and link targets are not carried over. Scanned pages are pictures without characters; for those you get an explanation instead of an empty .md.
This tool is also known as pdf to markdown, pdf to md, convert pdf to markdown, pdf to markdown for chatgpt, pdf to markdown converter online.
What is Markdown?
Markdown is a way of writing that gives plain text structure with a handful of simple marks: # at the start of a line means a heading, - a bullet, 1. a numbered item, and words between two asterisks are **bold**. The file is an ordinary text file with a .md extension; because the marks stay readable, it makes sense even without a renderer. GitHub README files, Obsidian and Joplin notes, many blogging platforms and the replies of AI chat assistants are written in Markdown. From a PDF this tool produces only the most common structural marks (headings, lists, bold); it writes no table, link or image syntax.
What is Text layer?
The text layer is the part of a PDF that says which letters are printed where on the page and in which font; if you can select a word in the PDF with your mouse, the document has one. Files exported from Word, Google Docs or a browser's print-to-PDF carry a complete text layer. A document that went through a scanner or a phone camera is only a picture of a page; there is no text to read until OCR (optical character recognition) has been run on it. PDF to Markdown conversion works on the text layer alone; when none is found, the tool warns you instead of producing an empty file.
How are heading levels chosen, and why are they sometimes wrong?
A PDF never says "this line is a level-2 heading"; the tool infers it from font size. It first finds the body size, the size in which most of the text is set. Lines at least 1.3 times larger and no longer than 120 characters are heading candidates; the distinct heading sizes are ranked from largest to smallest and the top three get #, ## and ###, with anything smaller staying at ###. Two consecutive heading lines of the same size are merged into one heading, because a long title is often broken over two lines in the PDF.
The blind spots are predictable. In a document whose headings are the same size as the body and only bold, no heading is found; those lines come out as **bold** paragraphs and you can add the # yourself. Big emphasis on a cover page, such as a date or a total, may be mistaken for a heading. Bold text in a font whose name does not reveal the weight (some embedded fonts hide it) is not marked bold. Table cells are taken row by row, left to right, and written out as ordinary lines; no Markdown table is built, because cell boundaries are not written in the PDF and a wrongly built table misleads more than plain lines do.
Markdown, plain text or Word?
If the document is headed for an AI model, a note-taking app or a website, Markdown is the right target: structure (headings, lists, emphasis) survives as a few characters while the file stays plain text that opens anywhere. Language models read Markdown well; a document split into sections is summarized more accurately than a PDF copy with lines broken at random.
If you only need the words (searching, counting, copying), the PDF to text tool gives a simpler .txt without any heading marks. If a person will edit and print the document, the PDF to Word tool writes the headings as Word styles. In short: Markdown when a machine will read it, plain text when you only need the words, Word when someone will edit and print it.
Frequently asked questions
Why convert a PDF to Markdown before giving it to ChatGPT or another AI?
Text copied straight out of a PDF breaks at the page width, page numbers and running headers end up between sentences, and headings blend into ordinary lines. After conversion to Markdown the paragraphs are whole, the repeated lines are gone and sections are marked with # headings, so the model can tell which sentence belongs to which section. On long documents this noticeably improves summaries and answers. The conversion runs in your browser; the PDF is not uploaded anywhere from this page.
How do I open a .md file?
A .md file is plain text: Notepad on Windows and TextEdit on a Mac open it, showing the marks (#, -, **) as they are. To see it formatted, use Obsidian, Typora, VS Code's preview, GitHub or Notion's import. On a phone, apps such as Markor (Android) and iA Writer render Markdown.
Do tables in the PDF become Markdown tables?
No. In a PDF a table is just separate pieces of text surrounded by lines; the cell structure is not stored in the file. The tool lays tables out left to right, row by row, as plain text. If you need the table, separate the cells in the output with | characters by hand, or copy the table into a spreadsheet.