Word frequency

Analyze how often words and characters appear in your text.

Unit Ignore case

Editors notice repetition long before readers can name it. A draft that leans on the same word nine times reads flat, and rereading rarely catches it — the eye slides right over the repeat.

Paste the text, run the count, and whatever you have been overusing sorts itself to the top.

How it works

Two units of analysis

The Unit menu decides what gets counted.

Words — the text is split on whitespace and on punctuation (period, comma, exclamation and question marks, semicolon, colon, brackets, quotes, backtick, tilde). Each remaining chunk is one token.
Characters (no spaces) — every character counts on its own with whitespace discarded. This is the mode that makes sense for Korean, Japanese, and Chinese text, and for typography or cipher work.

Ignore case folds the text to lowercase before counting, merging The and the into a single entry.

Reading the table

Press Analyze. The summary line gives the total token count and how many were unique; the table lists the top 30 by frequency with each token's count and its share as a percentage. Share is measured against every token in the text, not just the thirty on screen, so the visible percentages will not add up to 100.

Worth keeping in mind

There is no stop-word list, so the, a, and and will usually occupy the top rows in English — scan past them to the first content word, which is where a repetition problem actually shows. Tokens are not stemmed either, so run, runs, and running are counted separately. All of it is computed in your browser.

Terms explained

Token: One counted unit — a word or a single character, depending on the Unit setting.
Unique count: How many distinct tokens appear. A low unique count against a high total is a sign of heavy repetition.
Share: A token's count divided by the total number of tokens, shown as a percentage.
Case folding: Converting the text to a single letter case before counting, so that The and the are treated as one word.
Stop words: High-frequency function words such as the, a, and of. This tool does not filter them out.

Frequently asked questions

Why do 'the' and 'and' dominate the results?

Every token is counted with no stop-word filter, so English function words naturally take the top rows. Skip past them to the first content word — that is where genuine repetition shows up.

Do 'run', 'runs' and 'running' count as the same word?

No. Tokens are compared exactly as written, with no stemming or lemmatization, so inflected forms are counted separately. If you want them combined, normalize the forms before pasting the text in.

Why do the percentages not add up to 100?

Share is calculated against every token in the text, while the table shows only the top 30. The tokens below the cutoff hold the remaining percentage.

When should I use the character mode?

For Korean, Japanese, and Chinese text, where spaces do not mark word boundaries, and for any task where individual letters matter, such as checking which characters a typeface needs to cover.

Is there a length limit, and is my text uploaded?

Counting runs in your browser, so nothing reaches a server and there is no server-side limit. A very long document may make the page pause briefly while it counts.