How to Find Repeated Words in Your Writing

Word frequency is a descriptive tally. It lists how often each counted word appears in the text you already wrote. It does not grade style, and it does not predict search rankings. A word that appears often may be the subject of the piece, a necessary function word, or a habit you want to vary. The word frequency counter only shows the numbers so you can decide.

Repetition is not always a writing problem. Technical terms have to recur. A character’s name will dominate a scene. Instructions that say “turn” six times may be clearer than six synonyms. Use frequency as a flashlight, not as a score.

What word frequency measures

The analyzer uses the same word segmentation as the main word counter. Each word-like segment is lowercased, then added to a tally. Punctuation-only segments are ignored. The table shows the top 15 words, sorted by count descending. When two words share a count, they are ordered alphabetically so the ranking is stable.

Share of words is:

occurrences / total counted words × 100

That denominator is the full counted-word total even when you hide common function words from the table. The percentage would jump in a misleading way if the hidden words were also removed from the bottom of the fraction. The note under the table states which total is in use.

Repeated words

A repeated word is simply a token that occurs more than once after normalization. The tool does not know whether the repetition is elegant or dull. It cannot see that you used “river” as a theme. It can tell you that “river” appeared eight times in 400 words (2.0%).

Worked example. Paste:

The river followed the river path until the river met the lock.

Lowercased, the and river both repeat. If common words are shown, the likely leads. If you exclude common words, river rises in the visible ranking, but its share is still measured against every counted word in the sentence, including the. That is the documented choice so a hidden stopword list cannot inflate percentages.

Common function words

English leans on a small set of function words: articles, pronouns, auxiliaries, prepositions. They will dominate almost any natural paragraph. The checkbox “Exclude common words” removes a local, documented list of those items from the displayed ranking. It does not delete them from the editor and it does not change the share denominator.

Turn the box on when you want to see content words. Leave it off when you are studying how often you start sentences with there or it. The list is a convenience, not a linguistic authority. Domain language such as “data” or “patient” will never be on that function-word list, even if you repeat it.

Normalizing uppercase and lowercase

River, RIVER, and river count as the same word after lowercasing. That is usually what you want when you are looking for repetition. It also means a heading in capitals will merge with the same word in the body. If you need to preserve case for another task, use the case converter separately; frequency analysis always compares in lowercase.

Counting whole words

Lookup is a whole-word search against the same normalized tokens. Typing the does not increment for them or other. Typing art does not match start. That avoids a common trap in raw substring search, where a short query lights up inside unrelated words.

If lookup says a word does not appear, check spelling and whether the segmenter treated a hyphenated form as one token. well-known is typically stored as that whole token, not as well and known.

Reviewing frequently used vocabulary

After a draft exists, skim the top 15. Ask three practical questions. Is the leading content word the one you meant to be about? Are two near-synonyms splitting attention when one would do? Are filler verbs such as get and make doing work that a more specific verb could do once?

If the table is empty, the editor is empty. The page will say “No words to analyze yet.” It will not invent popular words to look busy.

Using frequency analysis during editing

Frequency updates are debounced so a long paste does not freeze the table on every keystroke. For ordinary typing, wait a moment after you stop and then read the ranking. For a 100,000-character paste, give the table a short processing window; the word and character totals in the margin still update promptly.

A useful loop: write, check the top content words, revise one paragraph, check again. If you cut repetitions of “very,” the share of remaining words will shift slightly because the denominator is the current counted total. That shift is arithmetic, not a quality score.

Second example. A product paragraph that uses “compact” four times in 80 words is 5% of the counted words. You might keep two and replace two with a measurement (“fits in a jacket pocket”) if the repetition feels idle. You might keep all four if compactness is the only claim that matters. The table will not choose for you.

Why repetition is not always a writing problem

Readers need recurring terms to stay oriented. Legal and scientific writing often requires the same noun so the referent does not drift. Children’s instructions should not rotate synonyms for “stop.” Even literary prose repeats a motif on purpose. A frequency counter that treated every repeat as an error would be a bad editor.

What the tool is good at is catching accidental clusters: the same adjective in consecutive sentences, a name spelled two ways that you can then unify, or a function word that starts every paragraph. Combine it with ordinary reading. If a sentence sounds fine aloud, a high count alone is not a reason to rewrite it.

Ties are broken alphabetically after count. If amber and river both appear four times, amber is listed first. That keeps the table from jumping around when you add a word that does not change those two counts. Sorting is a presentation rule, not a judgment about which word matters more.

For the underlying word rule, see how to count words in text. For cleaning space so tokens split the way you expect, see simple text formatting tools.

FAQ