How to use the word phrase counter
- Paste your text.
- Enter the number of words in each phrase, from 2 to 6.
- Set the smallest count to show, and choose whether to ignore capital letters and whether phrases may cross sentence ends.
- Read the ranked table of phrases with counts and shares. Download every phrase as CSV.
Formula
a word is a run of letters or digits, with an inner apostrophe kept
phrases = every run of n words in a row, sliding one word at a time
number of phrases = words - n + 1 when nothing is split
share = count of a phrase / number of phrases
When phrases may not cross sentence ends, each sentence is counted on its own, so the last words of one sentence are never joined to the first words of the next.
Repeated phrases in a short text
A text of 25 words about a quick brown fox has 22 two-word phrases when sentence ends are respected. The phrases brown fox, quick brown and the quick each appear 3 times, 13.64 percent of all phrases, followed by lazy dog with 2. There are 15 different phrases, and 4 of them are used at least twice.
With three-word phrases, quick brown fox and the quick brown lead with 3 uses each out of 19. If the phrases are allowed to cross sentence ends, the same text has 24 two-word phrases, 17 of them different, and brown fox still leads with a share of 12.5 percent.
What phrase counts show
Counting single words mostly finds small words such as the and of. Phrases are more telling: repeated pairs and triples show habits of style, standard wording and key topics. Writers use them to spot overused phrases, and researchers use them to compare texts.
Short texts give small counts, so a phrase seen twice may be chance. Longer texts give more reliable counts. Setting a minimum count of 2 hides the one-off phrases and keeps the table short.
Related tools
For single words, use the Word Frequency Counter. To find words repeated close together, use the Repeated Word Finder, and to see a phrase in its surroundings, use the Keyword in Context Concordance. To judge reading level, try the Readability Score Calculator.
Assumptions and limits
- Words are split on anything that is not a letter, a digit or an apostrophe inside a word.
- Sentence ends are a period, exclamation mark or question mark followed by a space, or a blank line.
- Text is limited to 200,000 characters and the table shows the top 50 phrases.
- Ties in count are listed in alphabetical order.
- The text and settings shown when the page opens are examples.
Frequently asked questions
What is an n-gram?
It is a run of n items in a row, here words. Two words in a row is a bigram and three is a trigram.
Why does the phrase count not equal the word count?
A text of w words has w minus n plus 1 phrases of n words, because the last few words cannot start a full phrase. Splitting at sentence ends lowers the count further.
Are numbers counted?
Yes, runs of digits count as words, so a phrase such as in 2025 can appear in the table.
Is my text kept?
No. The counting happens in your browser, and the text is not stored or sent anywhere.