How to use the text similarity checker to compare two texts
How to compare two texts
- Enter or paste content into Text A and Text B. Each field needs at least one non-whitespace character before the results appear.
- Select Upload Text A or Upload Text B to load a
.txt,.csv,.log, or.mdfile into the matching field. - Edit either text to recalculate the comparison. The result section stays hidden while either field is empty or contains only whitespace.
- Select Copy Report to copy the score, method, matched passages, and the two unmatched passage lists.
What the score and passage lists show
The score measures shared three-word sequences across both texts. Its label is Mostly distinct at 0 to 29%, Some overlap at 30 to 59%, Highly similar at 60 to 84%, and Near-duplicate at 85 to 100%.
Matched Passages pairs each Text A passage with the Text B passage that has its highest word-set overlap. Shared Content keeps the shorter passage from each matched pair. Unique to Text A lists Text A passages without a match, while Unique to Text B lists Text B passages that no match selected.
Text A: quick brown fox
Text B: quick red fox
Overall score: 0.0% (Mostly distinct)
Matched Passages: 50% quick brown fox | quick red fox
Shared Content: quick red fox
Unique to Text A: none
Unique to Text B: noneHow the checker calculates similarity
For the overall score, the checker lowercases the text and keeps Unicode letters and numbers as tokens. Punctuation is ignored. The tokens become overlapping three-word sequences, and cosine similarity between the two sequence-frequency sets produces the percentage. A text with one or two tokens uses those tokens as one sequence.
For passage results, blank-line paragraph breaks are read first. Sentence punctuation then splits each paragraph into passages; if no sentence boundary is detected, the paragraph remains one passage. Each Text A passage is compared with every Text B passage using word-set Jaccard overlap. A pair appears in Matched Passages at 50% overlap or higher, and one Text B passage can be selected for more than one Text A passage.
Where to use the comparison
Compare a revised article with an earlier draft, then inspect Matched Passages and Shared Content for wording that carried over.
For two summary versions, put one in Text A and the other in Text B to see which passages share vocabulary and which fall into the unmatched lists.
Generated text outputs can be compared, then saved with Copy Report for review of the score, pairings, and unmatched passages.
Choose the right text analysis tool
Text Similarity Checker belongs in Text Analysis Tools when the question is how much wording two texts share. Use Text Analyzer when you need counts, readability scores, lexical diversity, keyword frequency, or estimated reading time for one text. Choose Concordance Tool to find occurrences of a chosen word or phrase with surrounding context. Use Regex Tester when the task is testing a JavaScript regular expression against sample text.
Frequently asked questions
How does a text similarity checker calculate the overall score?
The overall score uses cosine similarity between overlapping three-word sequences across the two texts. Punctuation is ignored, and the percentage is rounded to one decimal place.
What does the text similarity checker match?
It matches each Text A passage with the Text B passage that has the strongest word-set overlap, provided that overlap reaches 50%. A Text B passage can match multiple Text A passages.
Why can similar text receive a low similarity score?
Similar meaning can receive a low score when the wording differs, because the checker compares tokens and overlapping word sequences. The checker has no semantic model and does not mark insertions or deletions.
Does the Text Similarity Checker send my text to a server?
The comparison runs in your browser with JavaScript, and the checker code does not send the entered text in a request.