Convert PDF to Markdown with the structure intact
The PDF to Markdown converter turns a PDF into tidy, readable Markdown: # headings worked out from the document’s font sizes, paragraphs with the line breaks of the page removed, **bold** and *italic* from the fonts, links from the PDF’s clickable areas, bulleted and numbered lists, and GitHub-flavoured tables for anything laid out in rows and columns.
Markdown is the native format of GitHub READMEs, Obsidian and Logseq vaults, Notion imports, static-site generators, documentation tools and AI chat assistants. Feeding a language model clean Markdown instead of a raw PDF keeps headings and tables recognisable and usually costs fewer tokens.
Conversion happens entirely in your browser, so research papers, contracts and internal reports are never uploaded.
How to convert a PDF to Markdown
- Drop a PDF onto the upload area or click to choose one. The whole document is converted straight away; type a range such as
2-6in Pages to convert only part of it. - Choose the options: Detect tables, Bold, italic and links, Skip page numbers, headers and footers, and Page breaks as ---.
- Check the rendered Preview, or open the Markdown tab to read and edit the source — the preview follows your edits.
- Press Copy Markdown or Download .md.
PDF to Markdown Converter features
Headings from font-size ranks
Every text size in the document is ranked: the largest becomes #, the next ##, then ###. A short bold line at body size becomes ####, so section titles in plain reports are kept too.
Real lists
Lines starting with •, –, *, 1., 2), a. or i. become Markdown list items, nested by indentation, and bullets drawn as tiny shapes rather than characters are recognised as well.
GFM tables
Ruled and borderless tables become pipe tables. Pipes in cells are escaped and numeric columns are right-aligned with ---:.
Inline formatting and links
Bold and italic come from the font names; links come from the PDF’s link annotations, even when only part of a line is clickable.
Cleaner text
Running headers, footers and page numbers that repeat on every page are dropped, words hyphenated across lines are re-joined, and characters that mean something in Markdown are escaped.
Live, sandboxed preview
The result is rendered with the open-source marked library in an isolated frame, and you can fix anything by editing the Markdown directly.
What each part of a PDF becomes
| In the PDF | In the Markdown |
|---|---|
| Largest text size | # Heading |
| Second and third largest sizes | ## Heading, ### Heading |
| Short bold line in body text size | #### Heading |
| Bold / italic font | **bold** / *italic* |
| Clickable link | [text](https://…) |
| • item, 1. item | - item, 1. item |
| Table | | A | B | pipe table |
| New page | Nothing, or --- with “Page breaks as ---” |
Getting the best Markdown from a PDF
- Digital PDFs convert best. Files exported from Word, Google Docs, LaTeX or a browser store real text and fonts. Scanned pages need PDF OCR first.
- Two-column papers are converted in the order the text is stored in the file, which is normally column by column.
- Headings all the same level? The document probably uses one font size throughout. Edit the
#marks in the Markdown tab before downloading. - Images are not included, because Markdown can only link to image files. Save them separately with a PDF image tool, or use PDF to HTML, which can embed them.
- To turn the Markdown back into a web page, use Markdown to HTML; for a printable file, Markdown to PDF.
PDF to Markdown Converter FAQ
How do I convert a PDF to Markdown?
Upload the PDF, check the preview, then press “Download .md” or “Copy Markdown”. The PDF to Markdown conversion is free, has no page limit and runs in your browser.
Does it keep tables?
Yes. Tables are detected from ruling lines and from text that lines up in columns, and written as GitHub-flavoured Markdown tables. Untick “Detect tables” if you would rather keep everything as paragraphs.
Which Markdown flavour is produced?
CommonMark with GitHub-flavoured tables, which works in GitHub, GitLab, Obsidian, Notion imports, Docusaurus, MkDocs and most other tools.
Is it good for preparing documents for ChatGPT or other AI tools?
Yes. Markdown keeps headings, lists and tables explicit, which helps a language model follow the structure, and it is far shorter than the raw PDF.
Why are some headings missing or at the wrong level?
Heading levels come from font sizes. If a document uses bold body text for headings, they become level-four headings; if it uses the same size and weight everywhere, there is nothing to detect. You can edit the result before downloading.
Is my PDF uploaded to a server?
No. The PDF is read by an open-source engine in your browser tab and the preview is rendered locally in a sandboxed frame.