Drop a PDF and get its word count, character counts, page count, and estimated reading time — computed on your device in seconds.
Translators quote per word, and clients send PDFs. Students face word limits, and the submission is a PDF. Editors, lawyers billing by document length, and writers checking a manuscript against a call for submissions all hit the same wall: PDF readers don't show a word count, and the workaround — select all, paste into a word processor — mangles formatting and takes minutes. This tool does it in seconds: drop the file, read the numbers.
The text layer of every page is extracted with pdf.js — Mozilla's PDF engine — in a background thread, and the counting happens instantly on the result: words, characters with and without spaces, pages, and an estimated reading time at a typical 200 words per minute. Character counts matter more than they used to; plenty of forms and abstract submissions limit characters rather than words, and per-character billing is standard in some translation markets.
A scanned document is a photograph of a page — there are no words in the file, only pixels shaped like words, so the count will honestly come back as zero. The quick test: if you can't select text in a normal PDF viewer, it's a scan. The site's OCR tools can turn scans into text; for digitally created PDFs — exported from Word, generated by software — the counts here are exact.
And because a word count usually gets requested for confidential material — unpublished manuscripts, legal documents, client files — the tool runs entirely in your browser. The PDF is parsed on your machine and never uploaded; there's no account and no cap on how many documents you check.
No. The text is extracted and counted by code running in your browser — manuscripts and contracts never leave your device.
It's almost certainly a scan: the pages are images with no text layer underneath. Run OCR on it first — this tool counts the embedded text only.
Very close but not always identical. Word and PDF extractors tokenize things like hyphenated words and numbers slightly differently, so counts can differ by a fraction of a percent.
Yes — everything in the page's text layer is counted, including running headers and page numbers, which is one source of small differences from the original word processor's count.