Keyword Density Analyzer
Analyze keyword frequency in any article — top words, 2-word phrases, and 3-word phrases.
Written by Anil Gautam · Updated
How does the keyword density analyzer calculate percentages?
The analyzer lowercases your text and splits it into words made of English letters or Devanagari characters, so digits, punctuation, apostrophes and hyphens all act as separators. Every row shows a count and a share. That share is the count divided by the total word count, and the total includes common words even while they are hidden from the lists. In a 400-word post where "home loan" appears 6 times, the phrase row reads 6 · 1.5%, because 6 divided by 400 is 0.015.
Phrase rows use the plain number of occurrences. Some SEO tools multiply a phrase count by the number of words in it, so compare figures from different tools with care. Each phrase also feeds the single-word list: those 6 mentions add 6 to "home" and 6 to "loan", and if "loan" appears 4 more times on its own, its row shows 10 · 2.5%.
What does the skip common words option change?
With the box ticked, 59 everyday English words, such as the, of, to, is, can and your, drop out of the single-word list, and any two- or three-word phrase that contains one of them is skipped. "Interest rate" stays, while "cost of living" disappears from the three-word list because "of" is one of the skipped words. Untick the box when the phrase you care about has a small word in the middle.
Under the text box, the words total counts everything and stays the same when you toggle the box. The unique figure counts distinct words after filtering, so it falls when common words are skipped.
Why do some words and phrases look odd?
Phrases are built from the running stream of words after punctuation is removed, so they can cross a full stop. Paste "Compare the best home loan. Loan rates change often" and the two-word list includes "loan loan", made from the end of one sentence and the start of the next. Give more weight to phrases that repeat several times.
Apostrophes and hyphens split words: "don't" becomes "don" and "t", and "e-mail" becomes "e" and "mail". An accented letter breaks a word too, so "café" is counted as "caf". Numbers are dropped, which turns "iPhone 15 review" into "iphone review", and singular and plural forms get separate rows. Hindi in Devanagari is counted, but a danda (।) typed straight after a word stays attached to it, so "है।" and "है" are counted separately. Tamil, Bengali, Gujarati and other scripts are not matched at all.
Paste only the article body, since menus and footers repeat on every page and push navigation words up the list. Each list shows its top 12 entries, with ties in the order they first appear. The analysis runs in your browser. To check the title and description that carry your main keyword, the SERP Preview Tool shows how they look in a results page.