How to convert a PDF to Markdown
To convert a PDF to Markdown, extract its text and let the tool infer structure: larger text becomes a # or ## heading, bold and italic runs become emphasis, and you get clean .md to copy or download. It works on text-based PDFs, never on a scan.
To convert a PDF to Markdown, open PDF to Markdown and let it read the page — free, without a signup, and in your browser with nothing uploaded. It pulls the text the PDF already stores, works out the structure from how that text is laid out, and hands you clean Markdown to copy or download as a .md file.
How to do it
Open PDF to Markdown
Go to PDF to Markdown. It runs on your device — no account, no install, nothing to upload.
Choose your PDF
Drop in a single text-based PDF. A password-protected file has to be unlocked first.
Convert, copy or download
Click Convert to Markdown. Every page is read, the Markdown appears in a preview with a Copy button, and a .md file downloads on its own. A note reports the page count and flags any page that had no selectable text.
Why convert a PDF to Markdown for an LLM?
Because Markdown is plain text that still carries structure, which is what a language model, a wiki or a notes app like Obsidian reads best. The headings and emphasis tell the model what is a title and what is body, without the byte weight and tag soup of a Word file or the awkward layout of the raw PDF.
How is this different from PDF to plain text?
PDF to Text joins the characters into lines and stops there — you get the words, with no structure. PDF to Markdown runs a second pass that reads structure back in: larger text becomes a # or ## heading and bold or italic runs become emphasis, so the shape of the document survives, not only its words.
What survives the conversion, and what does not?
Two heading levels survive, and two only: text at least 1.5 times the body size becomes a # heading, text at least 1.2 times becomes a ## subheading, and everything else stays a paragraph. Bold and italic are read from each font’s own name. Tables, links, images, columns and footnotes do not survive — the same limits as PDF to Word, because it is the same reconstruction underneath.
Does it work on a scanned PDF?
No — a scan has no text to read. The tool reads the characters a PDF already stores, and a scan is a photograph of a page with none underneath, so it stops and points you to OCR PDF rather than returning an empty file. Run OCR to add a text layer, as in how to make a scanned PDF searchable, then convert.
Is it private to convert a PDF to Markdown here?
Yes, because nothing is uploaded. pdf.js reads the pages and the Markdown is written in the same browser tab — there is no server to send the file to and no account to make. A confidential report, or a contract you are feeding to a local model, never leaves your computer.
The structure is inferred, not read from tags a PDF does not have, so it is a best guess worth skimming before you trust it — a heading missed here, an odd break there. Markdown’s own special characters in your text are escaped, so a literal # or * in the document renders as written instead of turning into stray formatting. If you only want the words with no headings at all, PDF to Text is the plainer path; if you want an editable document back, PDF to Word maps the same structure to a .docx.
