How to count words in a PDF
- Drop your PDF. There's nothing to configure.
- The text layer gets read. Page by page, on your device.
- Read the numbers. Words, characters with and without spaces, pages, words per page, and estimated reading time.
For the word limits that matter
Word counts matter when something is riding on them: a 2,000-word cap, translation billed by the word, a journal that specifies characters without spaces, which is why that figure sits alongside the rest here.
PDF viewers don't show a word count, and the obvious workaround, pasting the document into a counter website, means handing an unpublished draft to a server you know nothing about. This one reads the text layer where the file already lives. For the text itself rather than the tally, Extract text pulls it out to copy or save as .txt.
Why this number won't exactly match Word's
Put the same document through two counters and you get two answers. Neither is broken; they are counting different things, and if a hard limit is riding on the figure, it pays to know where the difference comes from.
- Everything on the page is text. A PDF has no idea which words are body copy and which are running heads. Headers, footers, page numbers, footnotes, table cells, figure captions and the address block on a letterhead all count here. Word, working from a document with real structure, can exclude several of those.
- Hyphenation at line ends. A justified PDF may break "unnecessary" across two lines. Whether that becomes one word or two depends on how the file marks the break, and it isn't always marked.
- Ligatures and unusual encodings. Some fonts store "fi" as a single glyph. Well-made PDFs map it back to two letters; poorly-made ones don't, which shifts the character count and occasionally the word count too.
- Text inside images. A chart with labels, or a scanned page, contributes nothing: those are pixels, not characters.
- Where the count was taken. A manuscript counted in Word before export includes tracked changes and comments that never made it into the PDF.
Expect agreement within a percent or two on a straightforward document, and a wider gap on anything with heavy furniture; journal articles with running heads and footnotes on every page are the usual outlier. When a limit is strict and close, count the text that actually matters: extract the relevant pages first, or pull the words out with Extract text, delete the furniture, and count what's left.
Reading time, and what it's worth
The reading-time estimate divides the word count by a typical silent reading speed for adult prose, somewhere around 200 to 250 words a minute. It's a reasonable guide for an article or a report someone will read straight through.
It is not a guide for anything dense. Technical documentation, contracts and academic papers are read at a fraction of that speed, with re-reading and cross-referencing that no word count can see. Nor does it describe reading aloud, which runs closer to 130 words a minute. If you are timing a presentation or a script, that is the figure to divide by, and the difference is large enough to matter over ten minutes.
Frequently asked questions
How do I count the words in a PDF?
Drop your PDF above. The text layer is read on your own machine, and back come the word count, characters with and without spaces, pages, words per page, and estimated reading time.
Why does the count differ from Microsoft Word's?
Counters disagree about what a word is. Hyphenated compounds, numbers, citations: every tool has its own rules, and the gap is usually under 1–2%. Against a hard limit, the number that matters is whatever your submission system reports.
Why does my PDF show zero words?
Zero means there's no text layer to read, which nearly always means the pages are scans. Run those through OCR first and they'll count normally. Exports from Word, Google Docs, or LaTeX count straight away.
Is my document uploaded to count it?
It stays where it is. Nobody should have to hand an unpublished manuscript to a stranger's server just to find out how long it is.