Free Keyword Density Checker

Keyword density is how often a word or phrase appears in a text compared with the total number of words. A phrase used 5 times in a 500-word article has a density of 1%. It is a simple count, and it says nothing on its own about whether the text is good or will rank.

The idea comes from early search engines, which leaned heavily on how often a page repeated the words someone searched for. Pages learned to repeat those words, and the number became a target. The habit of chasing a percentage stayed in many SEO checklists.

Today the count is still useful for two plain jobs: finding a phrase you repeated more than a reader would expect, and checking that the words you meant the page to be about really are the ones it uses most.

No. Google publishes no ideal keyword density: neither its SEO Starter Guide nor its spam policies give a density figure, and the Starter Guide does not mention density at all. Any percentage you see quoted as the right one is a rule of thumb from SEO writers, not a figure from Google's guidance.

What Google does name is the opposite problem. Its spam policies list keyword stuffing, which they describe as filling a page with keywords or numbers in an attempt to manipulate rankings, including repeating the same words or phrases so often that it sounds unnatural. The Starter Guide adds that repeating the same words over and over, even in variations, is tiring for readers.

That is why this checker shows counts and densities with no colored bands, no target range and no pass or fail. A density of 3% can be natural on a short product page about one item and forced in a long guide. Read the phrases at the top of the list and ask whether a person would write them that often.

Divide the number of times the phrase appears by the total number of words, then multiply by 100. The checker uses this one formula for every phrase length, so a two-word phrase that appears 4 times in 400 words has a density of 1%, the same as a single word that appears 4 times.

Keyword density = (times the phrase appears / total words) x 100.

Some tools count each word of a phrase separately, which makes a two-word phrase look twice as dense. Neither way is official. Counting occurrences per 100 words is easier to read: 1% means about once every 100 words, whatever the phrase length.

Words are counted the same way as in our readability checker and character counter, so the totals agree across the three tools. A word is a run of letters or digits, and it stays whole across an inner apostrophe or hyphen (don't, well-known) and across a decimal point between digits (3.5).

• Phrases never run across the end of a sentence, so "We sell boots. Trail maps too." has no phrase "boots trail".
• Matching ignores case and treats curly and straight apostrophes as the same, so Don't and don't count together.
• Matching keeps accents, because in each of the five languages an accent can tell two words apart: Spanish esta and está, French ou and où, German schon and schön.
• In French, an elided article or pronoun is dropped and the word after it kept, so l'outil counts as outil.
• The 30 most frequent phrases of each length are listed, most frequent first.

The target keyword is matched as whole words only, so "boot" does not match "boots" or "bootcamp". Its placement check tells you whether the phrase appears within the first 100 words, so you can see whether the opening names the topic.

Stop words are the short, frequent words such as the, of, and, und or de. In any text they are the most common words by far, so a keyword frequency list that kept them would fill its top rows with the and of and push the phrases you care about out of view.

The checker leaves out any phrase made only of stop words, such as "of the" or "und die". A phrase with at least one meaning word stays, so "history of the" and "price of" are counted. Pick the language the text is written in so the right list is used. Each list is the Snowball project's published stop word list for that language, and the tool links to each one with the date we last compared it.

Your target keyword is always counted in full, stop words included, because you asked for that exact phrase.

1. Paste the finished draft, not an outline, and choose its language.
2. Read the top of the one-word and two-word lists. If one phrase is far ahead of the rest and you would not say it that often out loud, rewrite some of the sentences that use it.
3. Enter your target keyword. If it never appears in the first 100 words, check that the opening says plainly what the page is about.
4. Look for the related words a reader would expect. A page about trail running that never uses shoes, distance or terrain is likely thin on the topic, whatever its density.
5. Read the edited text aloud. Google's spam policies name text that repeats words so often it sounds unnatural.

Replacing a repeated phrase with a pronoun, a shorter reference or a related term usually reads better and keeps the meaning. Removing it from the title or opening, where it tells readers what the page covers, usually does not.

In URL mode the checker fetches the page's HTML once, the way a crawler that does not run scripts reads it, and follows any redirects to the final address. It reads up to the first 3 MB of the page.

• The counts cover the main content: the article or main block of the page. Menus, sidebars, forms and the site header and footer are left out, so a menu link repeated on every page does not inflate a phrase.
• The title and the H1 are read from the whole page. The H2 to H6 headings and the image alt text are read from the main content, the same part the counts cover, so a sidebar or footer heading is not counted. Alt text is not counted as body words.
• A page that blocks our checker, answers an error or is not HTML is reported as such, with no counts.
• Text that a script adds after the page loads is not in the HTML, so it is not counted. If a page shows no words, it is probably built in the browser.