Keyword density checker
Paste an article, see the top 1-, 2-, and 3-word phrases — with percentages and a healthy-density warning.
HTML tags are stripped automatically. Stop-words are filtered for the single-word table only — phrases keep them so “how to write” surfaces correctly.
- Words
- 161
- Unique words
- 121
- Sentences
- 13
- Characters
- 983
| Term | Count | Density |
|---|---|---|
| ai | 4 | 2.48% |
| about | 3 | 1.86% |
| search | 2 | 1.24% |
| new | 2 | 1.24% |
| chatgpt | 2 | 1.24% |
| 2 | 1.24% | |
| answer | 2 | 1.24% |
| question | 2 | 1.24% |
| domain | 2 | 1.24% |
| seo | 2 | 1.24% |
| classic | 2 | 1.24% |
| llms | 2 | 1.24% |
| pages | 2 | 1.24% |
| answers | 2 | 1.24% |
| Term | Count | Density |
|---|---|---|
| ai search | 2 | 1.25% |
| the new | 2 | 1.25% |
| your domain | 2 | 1.25% |
| for the | 2 | 1.25% |
What keyword density actually means
Keyword density is the percentage of total words on a page that match a specific keyword or phrase. If a 1,000-word article uses the phrase “ai search visibility” eight times, the density of that phrase is 0.8%.
It's not a Google ranking factor in any direct sense anymore — Google's NLP figured out semantic relevance a decade ago. But the number is still useful as a writing diagnostic: if your primary keyword shows up at 5% density, you're probably stuffing. If a competitor ranks above you and their phrase frequency table is full of modifiers you forgot to include, that's a content gap to fix.
How to actually use this output
- Check your primary keyword. Healthy single-word density is 1-2% for your main term. Anything over 3% reads unnaturally; over 4% the tool flags it explicitly.
- Look at the 2-gram and 3-gram tables. These are the phrases real searchers type. If your article covers a topic but never uses the natural 2/3-word phrases for it, you're leaving relevance on the table.
- Compare to a top-ranking competitor. Paste their article in too. The phrases they use that you don't are content opportunities — not for keyword stuffing, but for genuine missing topics.
- Watch for accidental spam patterns. Repetitive 2-grams like “click here” or “buy now” stacking up in the table is a sign the article is template-y and conversion-y rather than substantive.
Why we filter stop-words for the 1-gram table
Without filtering, the top 5 single words on every English article on the planet are the same: “the, of, and, to, a”. Useless. The 1-gram table drops them so the most common content word floats to the top. The 2-gram and 3-gram tables keep stop-words because real phrases like “how to write” or “the best AI” need them to make sense.
FAQ
What's the ideal keyword density?
There isn't one. The advice you'll see online ('1-2% is the sweet spot') is folklore, not Google policy. Aim for natural phrasing — if you have to wedge the keyword in to hit a target percentage, the article will read worse and Google will notice.
Does Google penalize over-optimized articles?
Yes, but not via a 'keyword density meter'. Google's spam algorithms catch articles that read as keyword-stuffed regardless of the exact percentage — repetitive phrasing, awkward placement, content that doesn't match user intent. This tool flags 4%+ density as a smell signal, not because Google has a fixed cutoff.
Can I paste HTML directly?
Yes. We strip <script>, <style>, and all HTML tags before tokenizing, so a copy-paste from your CMS works fine. Inline text content (paragraphs, headings, list items) gets analyzed; navigation chrome and ads in the same paste get included too — paste only the article body for cleanest results.
Why are some words missing from the table?
We filter stop-words from the 1-gram table (the, and, of, etc.) so content words rise to the top. We also drop singletons (anything that appears only once) since one occurrence isn't a 'frequency'. Words shorter than 2 characters and pure numbers are also filtered as noise.
Does this work for non-English content?
Mostly yes. The tokenizer handles accented characters in European languages. The stop-word list is English-only, though, so non-English single-word tables will still be dominated by articles and prepositions in your language.
Will my pasted text be saved anywhere?
No. Everything runs in your browser. The text never leaves the page — that's why there's no rate limit.
Get the full Web Design Service workspace.
Want this monitored daily across all your projects? Sign up free.