Keyword Density Checker

Paste a block of copy to see how often each word recurs, how dense your focus term is, and which two-word phrases keep appearing. It counts every word for the length figures but strips common stop words out of the ranking, so the result reflects the words that actually carry meaning.

Counting Words, Then Counting the Ones That Matter

The checker reports two different kinds of number, and keeping them apart is the key to reading it well. Length metrics count everything, while the keyword table deliberately ignores a fixed list of common English words so that filler does not crowd out the terms you care about.

• Text is normalised before counting: it is lowercased, punctuation and symbols are removed, and the remainder is split on whitespace. So "Search-results" and "search results" both resolve to the same pair of words.
• The word and character counts use the normalised total, which includes stop words. The reading time wraps this up as one minute per two hundred words, rounded up, with a floor of at least one minute.
• The keyword ranking subtracts a built-in stop-word list covering articles, prepositions, pronouns and helper verbs — the, and, for, with, this, have and their relatives — so the table lists meaningful terms.
• Density is the term's count divided by the total word count, expressed as a percentage to two decimal places. Note the denominator includes stop words, which is why even a repeated keyword rarely exceeds a few percent.
• A focus keyword field lets you pin one term and get a labelled status: optimal between one and three percent, too high above three, and low below one.
• A second table lists the top two-word phrases, which catches repetition the single-word view misses, such as a phrase like "content creation" recurring across a page.

Reading a Result

1. Replace the sample paragraph in the Content box with the copy you want to test. The box accepts multi-line text, so pasting several paragraphs works.
2. Enter a focus keyword if you want a single term tracked and graded. The field is optional and matches on the normalised, lowercase form of the word.
3. Press Analyze. The summary tiles appear first, showing total words, total characters and estimated reading minutes.
4. If a focus term was supplied, read the strip below the tiles: it gives the occurrence count, the density percentage and the status word.
5. Scan the keyword table, which lists up to fifteen terms in descending order of frequency, each with its count, density and a bar relative to the most frequent term.
6. Check the two-word phrase table last. Repeated phrases often reveal a stylistic tic that single-word counts hide.

The Built-in Sample, Measured

The page ships with a five-sentence paragraph about search engine optimisation and sets the focus keyword to SEO. Running it produces a result that is worth reading closely because it shows the denominators at work.

• The sample contains forty-five words and three hundred and fourteen characters after normalisation, which the page reports as one minute of reading, since forty-five divided by two hundred rounds up to a single minute.
• The focus term SEO appears three times, giving a density of six point six seven percent. That is above the three percent ceiling, so the status reads Too high rather than Optimal — a reminder that short passages push densities up even when the absolute count is small.
• The frequency table is topped by SEO and content, both on three occurrences and both at six point six seven percent. Two terms — search and optimization — appear twice, at four point four four percent each.
• Everything else in the table sits at a single occurrence and two point two two percent. That long tail of one-offs is normal for prose and is not a signal by itself.
• The phrase table is dominated by one-off pairs such as "search engine" and "engine optimization". None repeats in the sample, so the phrase view offers no additional signal here — the repetition shows up only at the single-word level.

How to Judge Keyword Repetition

Alternative	Notes
Plain frequency, as here	Fast and transparent: it counts what is on the page. Its weakness is context — it cannot tell a keyword used naturally from one stuffed in, and its percentage depends heavily on document length, so short pages look dense even when the term appears only a few times.
Reading the draft aloud	The most reliable test for whether a phrase is overused is still to read it. Clunky repetition is audible. This catches the exact problem a density number is trying to approximate, without needing a target percentage at all.
Checking the search-result snippet	What a reader sees before clicking is the title and description, not the body density. Making sure the focus term appears there and that the description reads naturally often matters more than tuning the body count.
Covering related phrases instead	Rather than repeating one exact term, strong copy tends to use related phrasings — variants that carry the same meaning. A single-term checker will under-report that coverage, so pair a density check with a read-through for natural variation.

What Happens to the Text You Paste

The analysis is done entirely within the page, with no server round trip, and the text you paste is not transmitted anywhere. There is one local detail worth knowing if you are handling unpublished copy.

• The counting and ranking run in the browser as plain string operations, so the paragraph never leaves the machine and no network request carries it.
• The tool does record a compact history entry each time you analyse: the date, the word count, the focus term and its density. It deliberately does not store the pasted text itself.
• That history is kept in your browser's local storage under a key the tool owns, capped at the twenty most recent runs, and it disappears if you clear site data.
• Reloading the page returns the content box to the sample paragraph; the pasted copy is not restored.
• Because punctuation is stripped before counting, a result will differ slightly from the word count of a word processor, which typically treats hyphenated and punctuated forms differently. The tool's own count is the one the density figures are based on, so use it consistently.

Density Checker Questions

Why do the word count and the keyword percentages use different sets of words?

Because the percentage has to describe the whole document to be meaningful. The denominator is every word, including the and of and the. The numerator comes from a tally that omits stop words. So a keyword appearing twice in a two-hundred-word passage reads as one percent, even though the table that ranks terms is built only from the non-stop words.

What makes the status show Too high on such a short sample?

The optimal band is one to three percent, and it is a ratio, not a count. The sample paragraph is only forty-five words long, so three mentions of one term already reach six point six seven percent. On a longer article, the same three mentions would fall into the optimal band. Treat the label as a property of this document's length, not a verdict on the wording.

Does it count phrases as well as single words?

It lists the top two-word phrases in a separate table, but the density percentage applies only to single words. Any repeating pair is counted by sliding a window of two across the normalised word list and tallying the results, so overlapping pairs are all captured. Phrases are shown by count, without their own percentage.

Will it tell me whether my writing is any good?

No. It measures repetition, which is one narrow signal. It cannot judge readability, accuracy or whether a term is used naturally, and a low density is not automatically better. Use it to spot unintended repetition and to check that a focus term is present, then judge the draft by reading it.

Last verified: October 3, 2026

Launch Keyword Density Checker

All Tools · All Tools · About · How We Test · Contact · Privacy

Maintained by Huang Bin · Last updated October 3, 2026