Word Frequency Analyzer and Keyword Density Counter

Count how often each word, 2-word phrase or 3-word phrase appears in a text. Filter English stop words, set a minimum count, and copy the table or download it as CSV.

Runs in your browser: your files and text are not uploaded. Free, no account needed.

Last updated: · Reviewed by the QuickTooly Team

How Words Are Counted

How the Text Is Split into Words

  • Case folding: everything is lowercased first, so "Apple" and "apple" are one entry.
  • Kept characters: letters in any script (including accented letters such as "café"), digits, apostrophes (') and hyphens (-). Everything else, including other punctuation, becomes a space. "e-mail" and "don't" stay single tokens.
  • Curly apostrophes count as straight ones: "don’t" and "don't" are the same entry.
  • Edge symbols are trimmed: apostrophes and hyphens at the start or end of a token are removed, so "'hello'" counts as "hello" and a hyphen with spaces around it is not counted at all.
  • Words are found by spaces: Chinese, Japanese and Thai text, which does not put spaces between words, is counted as one long token per run of text. The stop-word list is English only.

Stop Words and Phrases

The stop-word list holds about 200 common English function words (the, and, of, it's, would and similar). With the filter on in single-word mode, these words are removed before counting, so they are left out of both the table and the total. In 2-word and 3-word modes the text is not shortened. Phrases are built from the full text, and a phrase is hidden only when its first or last word is a stop word. "Cost of living" is kept, while "of the" and "the cost" are hidden. Hidden phrases still count toward the total that the percentages are based on.

Reading the Percentages

Frequency % is the count divided by the total number of tokens (words or phrases) after any stop-word removal, so the same word shows a higher percentage with the filter on. "Unique words" counts only entries that meet the minimum frequency. Top 10, 25 or 50 limits the table only, and the chart shows the first 15 rows. Search engines do not publish a target keyword density, so use the numbers to spot repetition, not to hit a percentage.

Frequently Asked Questions

Does it work with accented letters and other languages?

Yes for languages that separate words with spaces. Accented and non-Latin letters are kept, so "naïve" and "café" count as whole words. Chinese, Japanese and Thai have no spaces between words, so each run of text counts as one token. Stop words are filtered for English only.

Why does the percentage change when I turn off stop words?

The percentage is based on the total token count. Removing stop words shrinks that total, so every remaining word gets a larger share.

What is in the CSV file?

Four columns: Rank, Word (or Phrase), Count and Frequency%. The file contains the rows currently shown, so set Top N to All first if you need every entry.

Related Text Tools