Text Similarity Checker
Score how similar two blocks of text are — a character-level similarity percentage plus a word-overlap breakdown showing shared and unique words.
Input
Treat "Word" and "word" as different words in the word-overlap breakdown.
Output
Character-level similarity (Sørensen–Dice coefficient over letter pairs).
| Metric | Value |
|---|---|
| No data yet | |
Guides
The Text Similarity Checker compares two blocks of text and scores how alike they are, both at the character level and at the word level. It's useful for spotting near-duplicate content, checking paraphrased text, or measuring how much two drafts, translations, or transcripts overlap.
What it measures
The main similarity score is a character-level comparison using the Sørensen–Dice coefficient: both texts are broken into overlapping pairs of letters, and the score reflects how many of those pairs the two texts share. It's sensitive to wording, spelling, and phrasing differences even when the overall meaning is similar.
The breakdown table adds a word-level view:
- Word overlap (Jaccard) — the percentage of unique words shared between the two texts, out of all unique words across both.
- Common words — how many distinct words appear in both texts.
- Unique to Text A / Text B — how many distinct words appear in only one text.
How to use it
- Paste the first text into Text A and the second into Text B.
- Toggle Case sensitive if "Word" and "word" should count as different words in the word-overlap breakdown (the character-level score is always case-sensitive).
- Read the similarity percentage and the breakdown table — use the copy or download button to export the breakdown as CSV.
How the score is calculated
Character similarity strips whitespace from both texts, splits each into overlapping two-letter sequences (bigrams), and calculates 2 × shared bigrams ÷ (total bigrams in A + total bigrams in B). A score of 100% means the texts are identical; 0% means they share no letter pairs at all. Word overlap instead treats each text as a set of words and calculates the size of their intersection over their union.
Common uses
- Editors and writers checking how much a revised draft changed from the original.
- Content teams spotting near-duplicate pages or descriptions before publishing.
- Students and researchers comparing paraphrased passages or translations.
- Developers deduplicating near-identical strings in a dataset.
Is my text private?
Yes. The comparison runs entirely in your browser — neither text is ever uploaded to a server.
Is this a plagiarism detector?
No. It measures textual overlap between two specific passages you provide — it doesn't search the web or any database for matching sources.