Word Frequency Counter
Paste any text and get a ranked table of what actually repeats — single words, or two and three word phrases — with a raw count and a percentage share for each. Useful for spotting the crutch word you have used nineteen times, and for seeing what a piece of writing is genuinely about rather than what you intended it to be about.
Use this without the search next time. Prathom Workbench puts Prathom's tools in your toolbar.
Add to Chrome — free| Term | Count | Share |
|---|
What it does
- Single word, two word, and three word phrase counts
- Percentage share alongside raw counts
- Optional filtering of common function words
- Full results exportable as TSV for a spreadsheet
How to use Word Frequency Counter
- 1
Paste your text
An article, a transcript, a page of marketing copy. The table builds as you type and re-ranks on every change.
- 2
Choose single words or phrases
Start with single words to find crutch words. Switch to two or three word phrases to see recurring constructions rather than vocabulary.
- 3
Filter if the table is noisy
Hide common words to drop function words like "the" and "of", or raise the minimum length to ignore short words entirely.
- 4
Export the full table
The page shows the top fifty terms. Copy as TSV to get every term, ready to paste straight into a spreadsheet column.
How it works
Text is first split into tokens by matching runs of letters and digits, optionally joined by an internal apostrophe or hyphen. That single rule is what keeps "don't" and "state-of-the-art" whole while still stripping the full stop off the end of a sentence.
Filters then apply in a fixed order: minimum length first, then common-word removal. Both operate on the token stream before any phrase is built, which is the decision described in the FAQ above.
For single words, the tokens themselves are counted. For phrases, a sliding window of two or three tokens walks the stream, joining each window into one term. A passage of n tokens therefore yields n − 1 two-word phrases, which is why the phrase totals are slightly smaller than the word total.
Counting is a hash map, then a sort by count descending with an alphabetical tie-break. The tie-break exists so that two runs over the same text always produce the same table — without it, the relative order of everything appearing once would be arbitrary.
What the numbers are good for
Crutch words. Every writer has them. Reading your own draft will not reveal that you opened eleven sentences with "Basically", because your eye slides over it. A ranked list will.
Unintentional emphasis. A term sitting at six or seven percent is being hammered, whether or not you meant to. This is the most common thing the tool surfaces on marketing copy — a product name or a single adjective repeated far past the point where a reader stopped noticing it as information.
Transcript summarizing. Two and three word phrases on a meeting or interview transcript pull out the recurring subjects surprisingly well, without any language model involved.
Variety as a rough signal. The unique-to-total ratio is a crude measure, but a sharp drop between two drafts of similar length usually means the second one started circling.
What it is not
This is a counter, not an analyzer. It has no notion of meaning, synonyms, or importance. A word that appears once can be the most important word in the document, and the tool will place it at the bottom of the table alongside everything else that appeared once.
It also cannot tell you whether repetition is a problem. Technical writing repeats terms deliberately, because substituting a synonym for a defined term is worse than the repetition. The table tells you what is happening; deciding whether it should be is still your job.
Examples
Single word counts on a short passage
Terms with equal counts are ordered alphabetically rather than left in whatever order they were found, so the table is reproducible. "Variety" is unique terms divided by total terms — a rough measure of how much a text repeats itself.
Two-word phrases reveal what single words hide
Counted as single words, "free" and "online" each appear twice and look unremarkable. Counted as a phrase, "free online" is a third of everything in the passage — which is the pattern a reader actually notices and the one worth editing.
Frequently asked questions
Is keyword density still worth optimizing for?
No. Keyword density has not been a meaningful ranking factor for well over a decade, and deliberately hitting a target percentage produces text that reads badly to both people and modern language models. The percentage column here is useful as a diagnostic — a term at eight percent is a sign you are repeating yourself — not as a number to tune toward.
Why is "the" always at the top?
Because English word frequency follows a steep distribution: a handful of function words account for a large share of any text, and "the" is the most common word in almost every English document ever measured. That is a fact about the language rather than about your writing, which is why filtering those words is offered as an option rather than applied by default.
Does removing common words change phrase results?
Yes, and in a way worth understanding. Phrases are built from the filtered stream, so removing "the" from "the quick brown fox" leaves "quick brown" and "brown fox" as phrases. That is usually what people want. But it also means words that were never adjacent can become adjacent after filtering, producing phrases that do not literally appear in your text. Check any surprising phrase against the original.
How are contractions and hyphenated words handled?
They stay intact. The tokeniser keeps apostrophes and hyphens that sit between letters, so "don't" and "well-known" each count as one term rather than being split into two. Leading and trailing punctuation is stripped, so a word at the end of a sentence is not counted separately from the same word mid-sentence.
Does it group singular and plural forms together?
No. "Tool" and "tools" are counted as separate terms, as are "run", "runs", and "running". Grouping them would require stemming, which is language-specific and frequently wrong in ways that are hard to notice — it would silently merge terms you wanted to see separately. Counting exact forms is predictable, and you can always add the numbers yourself.
Further reading
- Why Two Word Count Tools Give You Different NumbersWord count sounds like an objective measurement. Paste the same paragraph into Word, Google Docs, and a browser tool and you can get three different numbers. None of them is broken — they disagree about what a word is, and hyphens, numbers, and em dashes are where the disagreement lives.
- How Many Words Is a 5-Minute Speech?The short answer is 600 to 700 words, and 650 is the number to plan with. The longer answer matters more, because the reason most people overrun is not that they miscounted — it is that they used a reading pace to plan a speaking task, and those differ by about 30%.