Keyword Extractor

Pull the most frequent keywords and repeated two-word phrases from any text. The extractor counts frequency and runs in your browser.

AI Productivity Tools● Free No upload Instant

Loading Keyword Extractor…

Your browser is preparing the tool. It runs 100% locally.

Quick answer

Paste text and get its most frequent keywords: the top 15 words (three letters or more, common stop words removed) ranked by how often they appear, plus any two-word phrase that occurs more than once. It ranks by raw frequency, not TF-IDF, and your browser does the counting.

What the Keyword Extractor does

This finds the words and short phrases you use most. It counts every meaningful word (three letters or longer, with common stop words like 'the' and 'and' removed) and lists the top 15 by frequency, each with its count.

It also looks for repeated two-word phrases and lists up to eight that appear more than once. It is a frequency counter, not a language model, so it does not understand meaning.

How it works

Paste your text and run the extractor.

The tool lowercases the text, keeps words of three or more letters, and removes common stop words. It counts each remaining word and ranks them, showing the top 15 with their counts.

It also counts adjacent word pairs and lists those that repeat. Click any keyword or phrase to copy it. It all runs in your browser.

Generation algorithm

  1. Tokenise the text. The text is lowercased and split into words of three or more letters.
  2. Remove stop words. Common words such as the, and and of are dropped.
  3. Count and rank words. Each remaining word is counted; the top 15 by frequency are shown.
  4. Find repeated phrases. Adjacent word pairs that occur more than once are listed, up to eight.

How keywords are ranked

Keep words of 3+ letters; remove common stop words Rank by raw count (how often each word appears) Show the top 15 single words and up to 8 two-word phrases that repeat
Worked example
In text that repeats image 9 times and compression 6 times, image and compression top the list with those counts; a phrase like image compression appears if it occurs more than once.

Ranking is by raw frequency, not TF-IDF, so it rewards whatever you repeat most and cannot tell which words are distinctive.

Privacy

Your browser processes the text without a server or API call, so the text stays on your device.

Because it runs locally with no AI model in the cloud, even confidential text stays on your device.

You don't need an account and the site keeps no copy, so the result is gone once you close the tab.

Technical details

MethodWord-frequency counting
ModelNone (rule-based)
WordsTop 15 (3+ letters)
PhrasesRepeated pairs (top 8)
RankingRaw count, not TF-IDF
StemmingNone
Processing100% in-browser

Standards & references

  • Term frequency: ranks words by how often they appear. Stop words are removed because they are frequent but uninformative.
  • Not TF-IDF: TF-IDF also down-weights words common across many documents; this tool uses raw counts within your single text only.

Accuracy and limits

It ranks by raw frequency, not TF-IDF. A word that appears a lot rises to the top, even if it is not especially distinctive. TF-IDF, used by search engines, also weighs how rare a word is across documents. This tool does not.

There is no stemming. The words run, runs and running are counted as three different words, so related forms split the count instead of combining.

Only words of three or more letters are kept, and a fixed list of common stop words is removed. A meaningful short word, or one the list does not cover, may be dropped or slip through.

Phrases are limited to two-word pairs that repeat at least once; an important three-word term, or a phrase used only once, will not appear.

Where people use it

SEO keyword check

See which terms a page emphasises most, and how often.

Tag a document

Generate candidate tags or topics from an article.

Spot repetition

Find words you have leaned on too heavily in a draft.

Quick topic sense

Get a fast feel for what a long text is mostly about.

When it fits, and when it doesn't

Good for

  • A quick read of the most-used terms
  • Candidate tags or SEO terms from one text
  • Spotting overused words and repeated phrases

Not the best choice for

  • Distinctiveness ranking (use TF-IDF)
  • Combining word forms (no stemming)
  • Three-word or one-off phrases

For distinctiveness across many documents, a TF-IDF tool is better. To measure how dense a single term is, use the keyword density checker.

Frequently asked questions

How are keywords ranked?
By raw frequency: how often each word appears in your text, listed highest first with its count.
Is this TF-IDF?
No. It uses raw counts within your single text, not TF-IDF's weighting of how rare a word is across many documents.
Does it remove stop words?
Yes. Common words such as the, and and of are dropped before ranking, since they are frequent but uninformative.
Does it combine word forms?
No. There is no stemming, so run and running are counted as separate words.
How many keywords do I get?
The top 15 single words, plus up to 8 two-word phrases that appear more than once.
Why isn't an important phrase listed?
Phrases are limited to two words and must repeat. A three-word term, or a phrase used only once, will not show up.
What is the minimum word length?
Three letters. Shorter words are ignored, which also removes many small function words.
Is this AI?
No. It is a frequency counter with no understanding of meaning or context.
Can I copy a keyword?
Yes. Click any keyword or phrase chip to copy it.
Does capitalisation matter?
No. The text is lowercased before counting, so Image and image count together.
Is my text uploaded?
No. The Keyword Extractor runs in your browser and sends nothing.
How do I check one keyword's density?
Use the keyword density checker, which reports a single term's percentage of the text.

References

Counts how often each meaningful word appears (ignoring common stop words) and lists the top 15, plus any two-word phrase that repeats. It ranks by raw frequency, not TF-IDF, and does no stemming, so word forms are counted separately. Runs in your browser.

TrendingHot

Text Summarizer

Picks the 1 to 8 sentences that matter most. Rule-based.

AIOpen
Popular

Meta Description Generator

Trims your opening sentence to about 155 characters.

AIOpen
Trending

Paragraph Rewriter

Fixed plain-English edits that tighten wordy text.

AIOpen

Article Outline Generator

A fill-in blog post skeleton built around your topic.

AIOpen

Text Tone Analyzer

A rough positive, negative or neutral read from a word list.

AIOpen

Readability Analyzer

Check how easy a piece of text is to read.

AIOpen

Back to the Keyword Extractor

It's rule-based and runs in your browser, so your text stays with you. No account needed.