Updated · Built and checked by the PlinzoTools team, formulas tested against published sources
Calculate vocabulary richness with a type-token ratio (unique words vs. total words).
This tool counts how many distinct words appear in your text and calculates the type-token ratio -- the standard measure of vocabulary richness (or, conversely, repetitiveness).
Paste your text. Unique word count, total word count, and type-token ratio appear instantly.
Type-Token Ratio = (unique words / total words) × 100
The counter lowercases your text and extracts runs of letters, digits and apostrophes. Two forms of the same word in different capitalisation, such as "The" and "the", are therefore one unique word. Every word is included, common function words as well, so "the", "and" and "of" all count toward both the unique total and the overall total. Different word forms are not merged: "cat" and "cats", or "run" and "running", are two separate unique words. Hyphenated terms are split, so "well-known" contributes "well" and "known".
The pre-loaded sentence is "The cat sat on the mat while the dog watched the cat closely." It contains 13 words in total. The repeated words are "the" (four times) and "cat" (twice), so the unique words are the, cat, sat, on, mat, while, dog, watched and closely, which is 9. The type-token ratio is 9 divided by 13, or 69.2 percent. If you rewrote the sentence so that no word repeated, both totals would match and the ratio would be 100 percent. If you repeated "the cat" in every clause, the ratio would drop sharply. The tool shows the ratio to one decimal in its tile, while the copied summary uses two decimals.
A high ratio means a varied vocabulary, and a low ratio means more repetition. Neither is automatically good or bad. A children's story or an instruction manual deliberately repeats key words for clarity, so a lower ratio is fine. A creative essay or a set of product descriptions written to avoid sounding repetitive will score higher. In SEO work, a very low ratio in a long article can be a hint that a phrase is being overused, and it is a reason to read the piece critically, though it is not proof of a problem.
The ratio falls as texts get longer, because common words like "the" and "of" keep recurring while new vocabulary arrives more slowly. A 50-word paragraph can easily show 80 percent or more, while a 5,000-word chapter of the same style will show a much lower figure. For that reason you should compare texts of similar length only. If you want to compare a long article with a short one, take a section of matching size from the longer piece, for example the first 300 words of each, and compare those.
Because forms are not stemmed, the unique count can overstate how many distinct ideas a text contains: "analyse", "analysis" and "analysing" all count separately. Numbers count as words, so a text full of figures has a higher unique count without more vocabulary. The count also depends on how the text was cleaned, so pasting page furniture like menus adds words. Finally, a high ratio is not a synonym for quality. Deliberately swapping simple words for obscure ones to raise the score usually makes writing harder to read. Use the number to notice patterns, then judge them by reading.
Heavy repetition of the same words -- could indicate a narrow vocabulary or, in an SEO context, potential keyword stuffing.
Yes -- longer texts naturally tend toward lower ratios since common words repeat more, so it's most meaningful comparing texts of similar length.
No. The text is lowercased before counting, so different capitalisations of the same word are one unique word.
No. It compares exact word forms without stemming, so each variation is counted as its own unique word.
There is no single normal value. It depends heavily on text length and genre, so compare only texts of similar size and purpose rather than aiming for a fixed percentage.
No. Every word is counted, including common function words, in both the unique total and the overall total.
Reference: Google Search Central