What is a word frequency counter?
A word frequency counter tallies how many times each word appears in a text and ranks the results from most to least common. It turns "this draft feels repetitive" into a number you can act on, and it's the fastest way to see what a document is actually about. The top non-trivial words of an interview transcript or a competitor's page are its themes, whether the author intended them or not. This one counts in your browser, with stop-word filtering, a minimum-count cutoff, case-sensitive mode, and total/unique word stats. Nothing you paste is uploaded.
How to count word frequency with this tool
- Paste your text: a blog draft, transcript, survey responses, or a whole manuscript, up to 1 MB.
- Choose the analysis options:
- Exclude common English stop words: hides about 100 function words (the, and, of, to…) so the ranking shows meaningful vocabulary instead of grammar.
- Minimum count to include: set it to 2 or 3 on long texts to cut the tail of words that appear once.
- Case-sensitive counting: off by default (
Word=word); turn on to separate proper nouns from common nouns. - Show total and unique word counts: adds a two-line summary above the ranking.
- Read the ranked list. Each line is
word: count, sorted by frequency with ties alphabetical. Copy it out for a spreadsheet or report.
Example
Input:
The launch went well. The team shipped the update, and the update fixed the login bug.
Output with stats on:
Total words: 16
Unique words: 11
the: 5
update: 2
and: 1
bug: 1
fixed: 1
launch: 1
login: 1
shipped: 1
team: 1
well: 1
went: 1
Turn on stop-word filtering and the list starts at update: 2. Suddenly the ranking tells you the text is about an update, not about the word "the". That's the whole trick of word frequency analysis: raw counts measure grammar, filtered counts measure content.
Reading the numbers
"The" always wins, and by a lot. In ordinary English, the top-ranked word is usually "the" at roughly 5–7% of all words, the second word appears about half as often, the third a third as often, a pattern known as Zipf's law that holds across almost any natural text. This is why the stop-word filter exists: the top of an unfiltered list is nearly identical for every English document ever written.
Total vs. unique words measures variety. 16 total / 11 unique in the example above means each word is used ~1.5 times on average. Writers and linguists call unique-to-total the type-token ratio; if your 1,000-word draft has 250 unique words and a rival's has 400, yours is leaning harder on the same vocabulary. Compare like-sized texts only; the ratio always falls as texts get longer.
A high count isn't automatically a problem. A product page saying the product's name 14 times is fine; saying "innovative" 14 times is not. The counter finds the repetition; whether it's intentional is an editing decision. My own rule when editing drafts: any descriptive word in the top 20 that isn't the topic gets a synonym pass or a deletion.
What people use word frequency analysis for
- Editing: find your crutch words. Everyone has them ("really", "just", "actually" are common offenders) and they're invisible until counted. Run the draft, filter stop words, scan the top 30.
- SEO checks: see whether the term you want to rank for actually appears more than twice, and whether you've accidentally used a synonym twelve times instead. Ignore any advice about a "perfect keyword density percentage". No such number exists; the useful signal is which words dominate, not hitting a ratio.
- Research and feedback analysis: paste 200 survey answers and the top filtered words are your theme list ("pricing", "slow", "support") before you've read a single response. Same for interview transcripts and meeting notes.
- Language learning and teaching: frequency is the best predictor of which vocabulary to learn first; counting a real text you care about (song lyrics, a news article) shows which words unlock it.
Counting word frequency elsewhere
If your text already lives in another app:
| Where | How |
|---|---|
| Excel / Google Sheets | Put words in column A (Data → Text to Columns to split), then =COUNTIF(A:A, "word") per word; workable for a handful of words, painful for a full ranking |
| Python | from collections import Counter; Counter(re.findall(r"\w+", text.lower())).most_common(20), the standard approach for large corpora |
| Word / Google Docs | Ctrl+F shows a match count for one word at a time; neither builds a ranked list |
For a one-off ranked list from pasted text, the browser tool is faster than all three.
Quirks to know
Contractions split at the apostrophe
The tokenizer treats letters and digits as word characters, so don't counts as don + t. The stray t and s fragments are in the stop-word list, so filtering hides them; the don remains. For contraction-heavy transcripts, expand or strip apostrophes with the text replacer first if the exact counts matter.
Hyphenated terms count as two words
well-known becomes well + known. Consistent within a text, so comparisons still work, but remember it when a compound term seems undercounted.
Numbers count as words
2026 and 10 appear in the ranking. Usually harmless; raise the minimum count if a data-heavy text floods the tail with figures.
It counts single words, not phrases
New York counts as new + york. Frequent-phrase detection (n-grams) is a different analysis this tool does not do yet. For spotting repeated multi-word phrases, the duplicate word finder catches immediate repetitions, and a manual check of the top single words usually reveals the phrases behind them.
Related tools
- Word cloud generator - the same counts as a visual; big words = high frequency.
- Duplicate word finder - flags words repeated back-to-back ("the the"), which frequency ranking won't surface.
- Sentence counter - pair with frequency stats when editing for rhythm as well as vocabulary.
- Plain text converter - strip HTML before counting a scraped page.
- Remove line breaks - clean hard-wrapped PDF text before analysis.
Privacy and security
Counting runs entirely in your browser with no uploads and no storage. Paste unpublished manuscripts, user interviews, or internal docs without them leaving your machine.