Free Text Similarity Checker
A text similarity checker measures how alike two pieces of writing are. Paste both texts and this free tool returns a cosine similarity score based on the words each text uses, the share of vocabulary they have in common and the exact phrases that appear in both. It is handy for checking drafts, rewrites and near-duplicate pages. It compares only the two texts you paste.
Text Similarity Checker
Text and Content
About this tool
Cosine similarity turns each text into a list of word counts and measures the angle between the two lists. Identical word usage scores 100%, and texts that share no words score 0%. Because it looks at which words are used and how often, not their order, a text whose sentences were shuffled still scores high. The vocabulary overlap figure compares the sets of different words used, and the shared phrases show copying at the sentence level.
Common words like "the" and "and" appear in almost all English text, so by default the tool ignores a list of such stop words when scoring. That keeps two unrelated articles from looking similar just because both use small words. You can switch this off to compare raw word use, and choose how long a shared phrase must be, from two to six words. Longer phrases found in both texts are a stronger sign of reuse than short ones.
This is not a plagiarism checker. It never searches the web or any database, so it can only compare the two texts you paste. Use it to check a rewrite against its source, to see how close two product descriptions are, or to confirm that two pages on your own site are not near-duplicates, which can compete with each other in search results.
Frequently asked questions
What is cosine similarity?
It treats each text as a list of word counts and measures how closely the two lists point in the same direction. The result runs from 0% for no shared words to 100% for identical word use. It ignores word order, so it measures topic and vocabulary similarity rather than copied wording.
Is this a plagiarism checker?
No. It compares only the two texts you give it and never searches the internet. A high score tells you the two texts are alike, but it cannot tell you whether either one was copied from elsewhere. For that you need a service that checks against a large database of published pages.
What score counts as too similar?
There is no official threshold. Two rewrites of the same source often score above 70%, and unrelated texts on different topics usually score below 20%. For two pages on your own site, a very high score means they may compete for the same searches, so consider merging them or making each more distinct.
Why ignore stop words?
Words like "the", "of" and "and" make up much of any English text. If they are counted, two unrelated texts look more alike than they are. Ignoring them focuses the score on the meaningful words. You can tick the option off to include every word.
What are shared phrases?
They are runs of several words in a row that appear in both texts, such as "search engine optimisation services in". You choose the minimum length, from two to six words. Long shared phrases are a strong sign that one text was copied or lightly edited from the other.