All free tools Text analysed in your browser

Keyword Density Checker

See which words and phrases dominate a piece of content. Count words, estimate reading time, measure how often your target keyword appears and spot copy that repeats itself.

Analyse keyword density

Browser tool
Up to 300,000 characters. Pasted text is processed by JavaScript on this page and is never uploaded. Any language and script works, including Hindi.
Options

Measure density in four steps

Works for drafts that are not published yet and for live pages on any site.

  1. 1

    Add your text

    Paste a draft into the box, or switch to Fetch a page and enter a live URL to pull in its visible text automatically.

  2. 2

    Set a target keyword

    Type the phrase the page should rank for, such as windows vps. It is counted as a whole phrase and highlighted in the text.

  3. 3

    Analyse

    Select Analyse Text. You get word count, unique words, reading time and the top single words, two-word and three-word phrases with their density.

  4. 4

    Adjust the copy

    If the top phrases are menu text or your brand name, the content is thin on topic. If your keyword sits far above 3%, swap some uses for synonyms and related terms.

How density is calculated

Density is the number of times a word or phrase appears divided by the total word count, times 100. A 1,500-word guide that uses "managed vps" 18 times has a density of 1.2% for that phrase.

Matching ignores case, and phrases never run across a full stop, a line break or other sentence punctuation, so you only see phrases a reader would actually see together.

There is no perfect percentage

Search engines do not reward a fixed density. Pages that rank well usually mention their main phrase often enough to be clear, which tends to land between about 0.5% and 2%, alongside plenty of related terms.

Treat the numbers as a mirror. Placement matters more than repetition: the title, the H1, the opening paragraph, a subheading and the meta description.

Spotting keyword stuffing

One phrase repeated in nearly every sentence, a target keyword well above 3%, or three-word phrases that appear far more often than natural writing would produce are the classic signs of over-optimised copy.

Stuffed pages read badly and risk being treated as spam. Rewrite for the reader and let related vocabulary carry the topic.

Languages and scripts

Words are recognised in any Unicode script, so Hindi, Bengali, Tamil, Arabic or mixed Hinglish content is counted correctly. The stop word filter covers common English filler and frequent Hindi function words in both Devanagari and Latin spelling.

Reading time assumes about 220 words a minute, a typical pace for on-screen reading.

Privacy and limits

Pasted text is analysed entirely in your browser and is never sent to VPSLake. Only the optional URL fetch uses the server, and it returns the page text without storing it. The fetch reads the HTML the site sends, so text that a page loads later with JavaScript will be missing; in that case copy the text from your browser and paste it instead.

Keyword Density Checker FAQ

There is no official target. Write naturally, make sure the main phrase appears in the title, H1 and first paragraph, and use related terms throughout. If the checker shows your keyword above roughly 3%, the copy probably reads as repetitive.
No. Counting, phrase extraction and highlighting run in JavaScript on this page. You can disconnect from the internet after the page has loaded and the paste mode still works.
Scripts, styles, hidden elements, navigation menus, and the site header and footer are removed. When the page has a <main> element with real content, only that part is used, so the numbers reflect the article rather than the template around it.
As a whole phrase, in the same word order, within one sentence. Density is shown per occurrence, and for multi-word phrases the share of all words they occupy is shown too, since a three-word phrase used ten times takes up thirty words.
Tools disagree on what counts as a word, whether menus and footers are included, how hyphens and apostrophes are handled and which stop words are dropped. Compare figures from the same tool over time rather than across tools.
It works on public pages that return HTML. Pages behind a login, sites that block data-centre traffic, and single-page apps that render text only with JavaScript may return little or nothing. Private and internal network addresses are always refused.