📝

PDF to Markdown

Convert PDF text into structured Markdown — headings from font sizes, bullets preserved. 100% in-browser.

100% LOCALHeading heuristicsLists & italicsCopy or .md100% local

Saved Markdown· Auto-save on

▾

The CalculatorPage 1

📝

Drop a PDF here, or click to browse

Text-based PDFs (scanned pages need OCR first)

Insert page markers (--- Page n ---)

Drop a text-based PDF and convert it to structured Markdown

Data Source & Legal Disclaimer

Effective: 2026Last updated: 2026-09-07Update: Manual review

Sources: pdfjs-dist text layer

Markdown conversion runs entirely in your browser with pdf.js. Your file is never uploaded. Works on PDFs with embedded text — scanned pages need OCR.

See all data sources & update policy →

How it worksPage 2

↩ Back to calculator

How text becomes Markdown structure — illustrated

PDFs have no headings — only positioned glyphs. The converter groups each page's text items into visual lines by Y coordinate, then compares every line's font height to the page's median: ≥1.6× becomes #, ≥1.3× becomes ##, and ≤0.75× (footnotes, captions) becomes italics. Bullet glyphs (• – -) map to Markdown lists, numbered lines to ordered lists. This is honest heuristics, not magic: unusual layouts can mis-title a line, so skim the preview. If the output is empty, the PDF is scanned images — use the PDF OCR tool instead.

Text items → visual lines → font-size headings → Markdown

Font height relative to the page median is the only heading signal PDFs provide — bigger means more important.

Publishing a spec to the wiki — Sam's conversion

Sam has a 12-page PDF spec whose wiki import eats Markdown, not PDFs.

Convert:He drops the PDF; page 3's font-height scan finds the 18 pt section titles at 1.64× the 11 pt body median — they become ## headings, and the 24 pt document title becomes #
Inspect the heuristics:The preview shows one false positive — a large pull-quote became a heading; he fixes that single line by hand, far cheaper than retyping 12 pages
Count the structure:The stats row reads 4,180 words · 31 headings · 1 title — matching the spec's 12 numbered sections plus subsections
Import:spec.md downloads and the wiki renders headings, bullet lists, and paragraphs correctly on the first try

↩ Back to converter

FAQ & detailsPage 3

↩ Back to calculator

Convert PDF to Markdown online free. Everything runs locally; nothing is uploaded.

FreeToolHub PDF to Markdown is a free browser-based tool — convert PDF to Markdown online free. No signup, no upload; everything runs locally in your browser.

About this tool

What is this tool?

Convert PDF to Markdown online free. Font-size heading detection, lists, copy or download .md.

Heading heuristicsLists & italicsCopy or .md100% local

From PDF Layout to Markdown Structure

PDFs store positioned glyphs, not paragraphs; Markdown needs structure. The converter first rebuilds lines by clustering characters on shared baselines, then infers hierarchy from typography: lines at 1.6x the median font size become top-level headings, 1.3x becomes subheadings, and bullets become list items. The result is a Markdown file that reflects the visual hierarchy, not just the character stream.

Honest Heuristics, Editable Output

Font-size heuristics are usually right and occasionally surprised by letterheads or title pages - the tool demotes duplicated headings and defaults the document title to the file name to blunt the worst cases. Output lands in an editable text area: fix a heading if the heuristic guessed oddly, then copy or download. For scans with no text layer at all, the tool points you to PDF OCR instead of returning an empty success.

What Converts Well — and What Needs a Cleanup Pass

Text-based PDFs convert cleanly: headings come out as markdown headers, paragraphs stay paragraphs, and bullet lists survive as lists. Three sources need a human pass after conversion. Scanned pages carry no text layer at all — run the OCR tool on this site first, then convert. Complex multi-column layouts can interleave columns in reading order, so check any two-column page against the original. Tables convert as plain text rather than markdown tables, because PDF tables carry no structural markup — the lines exist visually, not in the data. For best results on reports and papers (the cleanest inputs), keep the page-break option on if you plan to cite specific pages, and off if you want one continuous document for an editor or a knowledge base import.

Frequently Asked Questions

How are headings detected?

The tool reads every character font size, computes the median, and classifies lines relative to it: at least 1.6x the median becomes a top-level heading, at least 1.3x becomes a subheading, and noticeably smaller text renders italic. This heuristic works across documents that use completely different point sizes.

What if the PDF has no text layer?

Scanned PDFs are images with no extractable text, and this tool will return an empty result rather than pretending. Use the PDF OCR tool instead - it recognizes text from page images and hands the result back as editable text.

Why does my output have an unexpected heading at the top?

Repeated large text (like a letterhead) can win the heading heuristic. The tool also demotes duplicate headings and uses the file name as the document title by default. A quick scan of the output before use is recommended - the heuristics are honest, not magic.

Does the conversion work on scanned PDFs?

No — scanned PDFs are photographs of pages with no extractable text. Run the PDF OCR tool on this site first to recognize the text, then convert. The converter is honest about this: a scanned page yields an empty or near-empty markdown section rather than made-up content.

Where does the conversion run — is my document uploaded?

Everything runs in your browser using pdf.js rendering and local text-extraction logic. The PDF never leaves your device, which makes the tool safe for contracts, internal docs, and anything confidential.