Paste your text and the sidebar ranks words by how often they appear, with counts and percentages. Detailed Metrics shows the number of unique words for a quick measure of vocabulary variety.
Finding words you overuse
Every writer has habits: really, just, actually, very, or a favourite adjective that appears in every paragraph. They are hard to spot when reading your own work and easy to spot in a frequency list. Turn off No Stop Words to include every word, or keep it on to focus on content words, then look for anything that appears far more often than its meaning justifies.
| Word | Count | Share | Action |
|---|---|---|---|
| perfect | 5 | 4.17% | Keep one, replace the rest with specifics |
| bag | 4 | 3.33% | Fine: it is the product |
| really | 3 | 2.5% | Delete: adds nothing |
Unique words and vocabulary variety
Repetition is not always a fault: technical terms, names and the subject of a page should repeat. Look for the words that carry no information.
The Unique Words row counts distinct words, ignoring capitals. Divide it by the total word count to get the type–token ratio: 300 unique words in a 600-word essay gives 0.5. The ratio falls naturally as texts get longer, because common words repeat, so compare texts of similar length only. A noticeably low ratio in a short text usually means repetition worth revising.
Repeated phrases
Switch the panel to 2-word or 3-word to rank phrases instead of single words. Repeated openings such as in order to, it is important or at the end of the day stand out quickly. In fiction, frequency lists also catch repeated character actions, such as how often someone nods or smiles.
Other uses
- Language learning: paste a text you are studying and learn its most frequent content words first.
- Survey answers and reviews: paste open-ended responses to see which words come up most.
- Transcripts: check how often filler words such as um, like or basically appear in a talk.
- Search content: confirm a page uses its topic words naturally, as described in the keyword density checker.
How words are counted
Words are split with the Unicode word rules built into your browser and lowercased before they are counted. Apostrophes inside words are kept, so don't is one word, and hyphenated compounds such as long-term are one word. Different forms of a word are counted separately: run, runs and running are three entries. The list shows the 50 most frequent terms.
The analysis runs entirely in your browser, so you can paste drafts, private notes or client documents.