Get the word count of a PDF document
The word count of a PDF is measured by extracting the document's embedded text layer and splitting it on whitespace — the same text you can select and copy in a viewer. This page does that entirely in your browser and reports the word count, character counts with and without spaces, an estimated reading time at 220 words per minute, and the 20 most frequent words with common stopwords removed. Scanned PDFs have no text layer and will count as zero until they are OCRed; the file never leaves your device.
How it works
Text in a PDF is not stored as flowing paragraphs. Each page's content stream positions individual glyph runs at coordinates — often word fragments or even single characters — and the mapping from glyph codes back to Unicode letters lives in per-font tables. This tool uses pdf.js to interpret those streams: it walks every page, collects the positioned text items in reading order, and joins them with spaces or line breaks depending on the geometric gaps between them. The result is the document's selectable text layer, reconstructed the same way copy-and-paste would see it.
Counting then happens on that extracted text. Words are the maximal runs of non-whitespace characters after normalization, so "state-of-the-art" counts as one word and a hyphenation split across a line break is rejoined where the extractor can detect it. Characters are counted twice — including and excluding spaces — because translation and typesetting work is often priced per character. Reading time divides the word count by 220 words per minute, a commonly used silent-reading average for adult non-fiction; a 5,500-word paper therefore shows 25 minutes.
For the frequency list, every word is lowercased and stripped of punctuation, then checked against a stopword list — the grammatical glue such as "the", "of", "and", "with" that would otherwise fill every top-10. What survives is ranked by occurrence, and the 20 most frequent remaining words are displayed with their counts, which is a surprisingly good one-glance summary of what a document is about.
The crucial limitation is honest and structural: all of this reads only the embedded text layer. A scanned PDF is photographs of pages — there are no glyphs to extract, and the count will be zero or near it. For scans, first run OCR on the PDF to generate a text layer, then count the OCRed copy.
Worked example: checking a journal submission limit
A researcher must keep a manuscript under 8,000 words for journal submission. Her 22-page, 1.4 MB PDF export goes into the tool; extraction takes about three seconds and reports 8,412 words, 52,344 characters with spaces, and a 38-minute reading time — roughly 400 words over the limit. The top-20 list is led by "sediment" (214), "cores" (156), and "holocene" (98), confirming the extraction picked up the real body text and not garbage encoding.
The count includes running headers and the reference list, which her target journal excludes from the limit. She exports a PDF of just pages 2–19 (body only) using a page-extraction tool, re-counts, and gets 7,650 words — comfortably under. Total time spent: about two minutes, with the unpublished manuscript never leaving her laptop.
Frequently asked questions
How do I check the word count of a PDF without Word?
Drop the PDF onto this tool: it extracts the document's selectable text in your browser and reports words, characters, reading time, and the most frequent terms in a few seconds. No conversion to Word or upload to a server is involved.
Why does my scanned PDF show zero words?
A scan stores each page as a photograph, so there is no text layer for any counter to read. Run the file through an OCR tool first to add a recognized text layer, then count that output.
Why does the PDF word count differ from Microsoft Word's count?
A PDF contains everything printed on the page — headers, footers, page numbers, captions, table text — while Word usually counts only the main story unless told otherwise. Hyphenation and line-break reconstruction also shift counts slightly, so differences of a few percent are normal.
How is the reading time calculated?
The word count is divided by 220 words per minute, a widely used average for adults reading non-fiction prose silently. Dense technical or legal material is often read at half that speed, so treat the figure as a lower bound for hard texts.
What are stopwords and why are they removed from the top-20 list?
Stopwords are high-frequency function words such as the, of, and, and to that dominate any frequency list without saying anything about the topic. Removing them lets the top-20 list surface the document's actual subject-matter vocabulary.
Is my document uploaded to count its words?
No — text extraction and all counting run in JavaScript on your own device, and the text itself never leaves your browser. That makes the tool safe for unpublished manuscripts and confidential reports.