Count words, characters, sentences, and paragraphs in any PDF — plus reading time, readability score, keyword frequency, and page-by-page analytics. No signup, no upload.
Drop your PDF to count words
Click to browse or drag & drop
TL;DR — Count words in a PDF in seconds
Four steps. Under a minute. Nothing ever leaves your browser.
Click Browse or drag and drop your PDF. Text extraction with PDF.js begins immediately, entirely inside your browser.
A full dashboard appears the instant extraction finishes — words, characters, sentences, paragraphs, reading time, and a readability score, with no clicks required.
Sort the page-by-page table by word count, search the extracted text with live highlighting, or look up a keyword to see occurrences and context snippets.
Download the full analysis as CSV, JSON, or TXT — ready for a report, an audit, or another tool.
Why PDFcrest
Eleven primary statistics, keyword search with context snippets, and per-page breakdowns — most free tools stop at a single total.
Words, characters (with and without spaces), sentences, paragraphs, pages, three averages, plus reading and speaking time — all computed the moment extraction finishes.
A genuine Flesch Reading Ease score and Flesch-Kincaid grade level — the same standard formulas used in publishing and government plain-language guidelines, not a made-up "easy/hard" label.
Search any word or phrase and get occurrence counts, a list of matching pages, and context snippets — not just a yes/no "found it."
See exactly which pages are longest or shortest, sorted in one click — useful for spotting where a document's content is concentrated.
Take the full statistics — including the top keywords and every page's numbers — into a spreadsheet, a script, or a report.
Your PDF is read directly in your browser's memory with PDF.js. It's never transmitted anywhere — after the page itself loads, the tool makes zero network requests.
Is This Tool Right For You?
A quick decision guide — this tool reads and measures text, it doesn't edit or restructure it.
Comparison
Adobe Acrobat and most free online counters show a single word total and nothing else.
| Feature | PDFcrest | Adobe Acrobat | Typical Free Online Tools |
|---|---|---|---|
| PDF Word Count — Free | ✓ Always free | ✗ Full Acrobat subscription | ~ Often free, some cap daily use |
| No Signup Required | ✓ Zero signup | ✗ Adobe account required | ~ Varies by tool |
| File Ever Uploaded to a Server | ✓ Never — runs in your browser | ✗ Uploaded to Adobe's servers | ✗ Usually uploaded to a server |
| Character, Sentence & Paragraph Counts | ✓ All included | ✗ Not shown | ~ Rarely more than word count |
| Reading & Speaking Time Estimates | ✓ 3 reading speeds + speaking | ✗ Not offered | ✗ Rarely offered |
| Readability Score (Flesch) | ✓ Built in | ✗ Not offered | ✗ Rarely offered |
| Keyword Frequency Table | ✓ Built in, stopword filter | ✗ Not offered | ✗ Not offered |
| Page-by-Page Word Count Table | ✓ Sortable | ✗ Not offered | ✗ Not offered |
| Export Statistics (CSV/JSON/TXT) | ✓ All three formats | ✗ Not offered | ~ Some offer basic export |
| Mobile Support | ✓ Any modern mobile browser | ~ App or desktop only for full features | ~ Varies |
Comparison reflects each tool's general publicly documented free-tier behavior at time of writing and may change — always check the provider's current terms.
Use Cases
Anyone who needs to measure, verify, or search the text inside a PDF relies on a PDF word counter.
Check a manuscript or article against a publisher's word count requirement before submission, without re-opening the original document.
Verify an essay, thesis chapter, or paper meets a professor's or journal's length requirement, and search for how often a key term appears.
Check the Flesch reading level before print, compare chapter lengths in the page table, and confirm estimated reading time for back-cover copy.
Search a contract or filing for every occurrence of a defined term or clause reference, with the exact page and surrounding context for each match.
Get an accurate source word count to quote a translation job, without manually copying text out of the PDF into a word processor first.
Audit a PDF resource or whitepaper's content depth and top keywords before repurposing it into a web page or blog post.
Confirm a report or public notice meets a required length or plain-language readability target before it's published or filed.
Workflow Recipes
Counting words is rarely the only step. Here are the tool combinations PDFcrest users chain together most often.
A scanned PDF has no text layer to count. Run OCR first to make it searchable, then come back here for accurate statistics.
Found out you're 500 words over the limit? Convert to an editable Word file and trim the content down.
Extract each chapter or section into its own file first, then run word counts on each part separately for a per-section breakdown.
Use the keyword frequency table to see what a document is actually about, then make sure the PDF's title and keywords metadata match.
Analyze what's inflating a PDF's file size, then analyze its actual text content — a complete before-you-publish check.
Need the text itself, not just statistics about it? Use PDF to Word to get fully editable content.
Technical Detail
A PDF page doesn't store "paragraphs" or "lines" — it stores a flat list of positioned text runs, each with its own x/y coordinates. This tool uses PDF.js, the same open-source rendering engine behind Firefox's built-in PDF viewer, to read that list via its getTextContent() API. Runs that share roughly the same y-coordinate are joined into one line; a vertical gap larger than about 1.7x the page's typical line spacing is treated as a paragraph break rather than a simple line wrap. That reconstruction is what turns a flat list of text fragments back into readable lines and paragraphs.
A "word" here is any run of Unicode letters or digits that may contain an internal apostrophe or hyphen — this correctly counts don't and well-known as one word each, and works on non-English scripts, not just ASCII. It does not treat punctuation as part of a word, so an email address or URL splits into its letter/number segments the same way most word processors count them, and a value like $3.50 counts as two number tokens rather than one currency value. These are deliberate, documented tradeoffs of a simple, fast tokenizer — not silent inaccuracy.
Sentence counting is rule-based: it protects common abbreviations (Mr., Dr., e.g., U.S., and about two dozen others), decimal numbers, and ellipses from being misread as sentence endings, then splits on a period, question mark, or exclamation point followed by whitespace and a capital letter, digit, or quotation mark. Paragraph counts come directly from the line-reconstruction step above — each block separated by a detected paragraph break is one paragraph.
The reading-level score is the standard Flesch Reading Ease formula, alongside a Flesch-Kincaid Grade Level — the same formulas referenced in U.S. federal plain-language guidance for public-facing writing. Both need an estimated syllable count per word, computed with a vowel-group heuristic (the same general approach standard readability libraries use). This heuristic is tuned for English; documents in non-Latin scripts fall back to one syllable per word, so the score is presented as an English-text estimate, not a universal metric.
Small differences between word counters are normal and expected — Microsoft Word, Google Docs, and PDFcrest all use slightly different rules for what counts as a "word" (hyphenated compounds, numbers with decimals, and stray formatting characters are the most common sources of disagreement). A difference of a few words per thousand is a tokenization-rule difference, not an error in either tool.
Glossary
PDF (Portable Document Format): a file format that preserves a document's exact layout, fonts, and images regardless of the device or software used to open it.
Word Count: the number of distinct words in a text, typically counted as whitespace- and punctuation-separated runs of letters and digits.
Character Count: the total number of characters in a text, usually reported both including and excluding spaces.
Sentence Count: the number of sentences, detected by punctuation that ends a sentence (., !, ?) while excluding abbreviations and decimal numbers.
Paragraph Count: the number of distinct text blocks, typically separated by a blank line or a significant vertical gap on the page.
Reading Time: an estimate of how long a document takes to read silently, based on an average words-per-minute reading speed.
Speaking Time: an estimate of how long a document takes to read aloud, based on an average spoken words-per-minute pace.
OCR (Optical Character Recognition): the process of detecting text within a scanned image and converting it into selectable, searchable characters.
Searchable PDF: a PDF containing an actual text layer that can be selected, copied, and searched — as opposed to a scanned image with no underlying text.
Unicode: the international standard for representing text in virtually every writing system, allowing accurate word and character counting in any language.
Text Extraction: the process of reading the underlying text content out of a PDF's internal structure, rather than treating the page as a flat image.
Keyword Frequency: a count of how many times each distinct word appears in a document, typically ranked from most to least common.
Lexical Diversity: the ratio of unique words (vocabulary size) to total words — a higher ratio means less word repetition.
Document Analytics: the broader set of statistics and insights — length, structure, readability, keyword patterns — that describe a document beyond its raw word count.
Under the Hood
No black box — this mirrors the real tokenizer, sentence splitter, and readability formulas that run in your browser the moment your PDF finishes extracting.
Word tokenizer — real regex, Unicode-aware:
/[\p{L}\p{N}][\p{L}\p{N}'’\-]*[\p{L}\p{N}]|[\p{L}\p{N}]/gu
Flesch Reading Ease:
206.835 − 1.015 × (words / sentences) − 84.6 × (syllables / words)
Flesch-Kincaid Grade Level:
0.39 × (words / sentences) + 11.8 × (syllables / words) − 15.59
Reading speeds — words per minute:
Silent Reading = 238 wpm
Presentation (aloud) = 130 wpm
Speed Reading = 450 wpm
Speaking Time = 150 wpm
| Component | What it actually does |
|---|---|
| Word regex | Matches a Unicode letter/digit run that may contain an internal apostrophe or hyphen — works across non-English scripts, not just ASCII. |
| Abbreviation list | ~28 common abbreviations (Mr, Dr, e.g., i.e., U.S., Inc., etc.) are protected from being misread as sentence endings before the sentence split runs. |
| Paragraph gap threshold | A vertical gap greater than 1.7× a page's median line-to-line spacing is treated as a paragraph break, not just a line wrap. |
| Syllable counter | A vowel-group heuristic feeds both Flesch formulas — tuned for English; other scripts default to one syllable per word. |
| Stopword filter | ~120 common English words (the, and, of…) are excluded from the keyword frequency table unless explicitly re-enabled. |
| Scanned-PDF detection | If total extracted non-whitespace characters fall under 10 across the whole document, the tool shows the OCR notice instead of a misleading all-zero dashboard. |
FAQ
Answers to the most common questions about the PDF Word Counter tool
References & Further Reading
Page last updated: · Written and maintained by the PDFcrest team. The counting logic shown above mirrors the real code running on this page.