PDF word counter

How it works

1. 1

   Drop your PDFs

   One file or up to twenty at once: a thesis, a contract, a stack of reports. Each can be up to 50MB.

2. 2

   Your browser reads the text

   The PDF is opened on your device with PDF.js, the same engine Firefox uses to show PDFs. Nothing is uploaded.

3. 3

   Get the full breakdown

   Words, characters with and without spaces, pages, words per page, reading time and tokens, with a row for each file.

What gets counted

A PDF stores its words in a hidden text layer that sits behind what you see on the page. The counter reads that layer page by page and counts every run of characters between spaces as one word, the same basic rule Word and Google Docs use. Numbers, URLs and hyphenated words each count once.

Everything in the layer is included: running headers and footers, page numbers, captions, footnotes and table cells. For an essay or manuscript with a strict limit, that page furniture can add a few words per page, so check whether your limit counts it. The words-per-page figure helps here: a page of dense prose usually lands between 400 and 600 words.

Scanned PDFs and empty pages

If a PDF came off a scanner or a phone camera, its pages may be pictures with no text layer at all. Many word counters report zero words for these, which looks like an answer but isn't one. This counter checks how much text each file actually has, flags files that look scanned and leaves them out of the totals. If only a few pages are images, it counts the rest and tells you how many pages had no text.

To count a scanned document, run it through OCR first. That adds a text layer, and the new copy can be counted here like any other PDF.

Words, reading time and tokens

Reading time assumes 238 words per minute, the average silent reading speed for non-fiction found in a 2019 review of 190 studies by Marc Brysbaert. Technical or legal text reads slower.

If the PDF is headed for ChatGPT or Claude, tokens matter more than words. The OpenAI count here is exact for current models; for older models and pasted text, use the token counter. A long PDF that runs past the context window gets searched instead of read in full, so splitting it into focused, clean files helps. The free knowledge base builder does that for a single PDF.

Frequently asked questions

How do I count the words in a PDF document?

Drop the PDF on this page and the count appears in a few seconds. Most PDF readers, including the free Adobe Acrobat Reader, don't show a word count, so the usual workaround is converting the file to Word first. This page skips that step and reads the PDF's text directly in your browser.

Why is my count different from Microsoft Word or Google Docs?

Usually by a few percent. This counter includes everything in the PDF's text layer, so headers, footers, page numbers and footnotes are counted on every page, while Word only counts what survives the conversion, which can merge or drop text. Like Word, a word here is anything between spaces, so "e-mail" and "3.5" are one word each, and a word that a line break split at a hyphen is joined back together and counted once.

Is my PDF uploaded anywhere?

No. The file is opened and counted inside your browser tab. It isn't sent to our servers or anyone else's, and nothing is stored, so it is safe for contracts, manuscripts and other private documents. Once the page has loaded, it keeps working offline.

It says my PDF looks scanned. What does that mean?

A scanned PDF is a set of page photos with no text layer, so there are no words for any tool to count. Rather than show a misleading zero, the counter tells you which files are affected. Run them through OCR (text recognition), for example Adobe Acrobat's Scan & OCR tool, or open the file with Google Docs from Google Drive, then count the new copy. If only some pages are images, the counter says how many were skipped.

Can I count words in several PDFs at once?

Yes. Drop up to 20 PDFs together, or add more later. You get combined totals plus a per-file table with the pages, words, characters and tokens of each document, which is handy for checking a batch against a word limit or a reading list.

How many words of a PDF can ChatGPT or Claude read?

Models measure text in tokens, not words. For ordinary English, one token is roughly three-quarters of a word, so 1,000 words is about 1,300 tokens. That's why the counter shows the exact OpenAI token count and a Claude estimate next to the words, and checks them against common context window sizes. Tables, code and non-English text use more tokens per word.

More free tools

All free tools →

Related guides