Word and Character Frequency Analyzer
Count word, character and phrase frequency in text with a ranked chart, stop-word filtering, and the caveats that make raw counts misleading.
With "skip common words" on, the phrases may never have occurred
Stop words are removed from the word list first, and the phrases are built from what is left. Removing a word therefore joins its neighbours together.
| Text | After filtering | Phrases reported | Not in the text |
|---|---|---|---|
| the rate of return on the fund | rate return fund | rate return, return fund | 2 of 2 |
| a bird in the hand is worth two in the bush | bird hand worth two bush | bird hand, hand worth, worth two, two bush | 3 of 4 |
| state of the art design | state art design | state art, art design | 1 of 2 |
| the quick brown fox jumps over the lazy dog | quick brown fox jumps over lazy dog | quick brown, brown fox, fox jumps, jumps over, over lazy, lazy dog | 1 of 6 |
On the first row every phrase is an invention. "rate return" is not something anybody wrote — it is "rate of return" with a hole in it. The cause is the ordering rather than the filter: turn the filter off and every reported phrase really does occur, by construction. Three-word phrases invent at least as freely as two-word ones, since there are more gaps to close. Text with no common words in it is unaffected either way.
This is the usual behaviour for a keyword tool, and it is usually what you want — it surfaces the topic rather than the grammar. It is worth knowing before quoting one of these as a phrase somebody used. If you want the phrases that genuinely occur, leave the 70-word filter off and read past the first few rows: "of the" will be at the top and the real phrases will be underneath it.
And the sentence count is really a punctuation count
Sentences are found by looking for runs of text ending in a full stop, question mark or exclamation mark. Abbreviations and decimals match that pattern too:
| Text | Counted | Actually |
|---|---|---|
| One. Two. Three. | 3 | 3 ✓ |
| Dr. Smith went home. | 2 | 1 |
| It cost $1.50 in total. | 2 | 1 |
| Wait... really? | 2 | 2 ✓ |
| No terminator here | 1 | 1 ✓ |
Ordinary prose comes out right; a title or a price adds one each time. The error only ever runs high, never low, because every real ending is also a match. So the bias always points the same way: prose with initials, prices or abbreviations reads as having shorter sentences than it has. The extreme case: "Dr. Smith and Mr. Jones met at 9.30 a.m. to discuss it." is one sentence of 14 words and counts as 6, giving 2.3 words per sentence instead of 14. Which matters if you are using words-per-sentence as a stand-in for difficulty, since it will make the text look easier than it reads.
How to use
- Paste the text you want analysed.
- Choose whether to count words, characters or phrases.
- Filter out common stop words to see the meaningful terms.
- Read the ranked results and the chart.
Frequently asked questions
What are stop words?
Very common words like the, of, and and to that dominate any frequency count while carrying little meaning. Filtering them out is what turns a word count into something that actually reflects the subject matter. There is no single authoritative list, so results vary slightly by which one is used.
Should I count word stems together?
Often yes. Treating run, runs and running as one word gives a truer picture of how much a text discusses running. Doing it well is harder than it sounds — crude stemming turns university into univers — so simple frequency counting usually leaves them separate.
What is Zipf's law?
The observation that in natural language, the second most common word appears about half as often as the first, the third a third as often, and so on. It holds remarkably well across languages and text types, and a frequency count that badly violates it usually indicates unusual or non-natural text.
What is phrase frequency useful for?
Spotting repeated constructions, filler and unintentional repetition in your own writing, and identifying key terms in someone else's. Two- and three-word phrases often reveal more than single words, since the meaningful units of a subject are frequently compounds.
Does capitalisation affect the count?
It does unless you fold case, which is usually what you want — The and the are the same word. The exception is when capitalisation is meaningful, as with proper nouns, where folding case merges a name with an ordinary word.
Can this tell me if two texts share an author?
Not reliably. Genuine stylometry uses function-word frequencies, sentence-length distributions and other subtle features, with statistical testing behind it. A raw frequency count is a starting point for that kind of analysis, not a conclusion.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.