Skip to content
TextArray
100% local

Text similarity

Measure how similar two texts are — similarity percentage, Levenshtein distance and word overlap.

Input
Second text
Output
Related tools
Text diffBasic tools
Compare listsLists & combinations
Word frequency counterBasic tools
JSON diffDeveloper tools
Pangram checkernewBasic tools
Word and character counterBasic tools
Text similarity
Paste two texts and get a similarity report on two levels. At the character level you see the Levenshtein distance — the number of single-character insertions, deletions and substitutions needed to turn one text into the other — and a similarity percentage derived from it: 100 % means identical, 0 % means nothing in common. At the word level the report shows the Jaccard similarity of the two word sets, plus how many unique words the texts share and how many appear only in one of them.
The two levels answer different questions. Character similarity is right for near-duplicates: two versions of a paragraph, a corrected translation against the original, product descriptions that were copy-pasted and lightly edited. Word overlap is right when order and phrasing differ but the vocabulary should match — checking whether two articles cover the same ground, or how much a rewrite actually changed.
"Ignore case" is on by default so that "Hello" and "hello" don't count as a difference; switch on "Ignore punctuation" to compare the words alone. For very long texts there is a safety limit: character metrics are computed on the first 5,000 characters of each text (the report says so when it happens), because the edit-distance calculation grows with the product of both lengths. Word metrics always use the full texts.
Both texts are compared entirely in your browser — nothing is uploaded, so contracts, manuscripts and unpublished copy are safe here.
FAQ
How is the similarity percentage calculated?
It is 1 minus the Levenshtein distance divided by the length of the longer text, shown as a percentage. Identical texts score 100 %; texts with no characters in common score 0 %.
What is the Levenshtein distance?
The minimum number of single-character edits — insertions, deletions or substitutions — needed to turn the first text into the second. "kitten" to "sitting" takes 3 edits.
What does the Jaccard word overlap mean?
The unique words shared by both texts divided by all unique words across them. It ignores word order and repetition, so it measures shared vocabulary rather than shared phrasing.
Is there a length limit?
Character metrics use the first 5,000 characters of each text — beyond that the edit-distance calculation would freeze the tab, and the report notes the truncation. Word metrics always use the full texts.
Is my text uploaded anywhere?
No. Both texts are compared entirely in your browser and never leave your device.