Levenshtein Distance Calculator

How it is calculated

The strings are split into Unicode characters, so an accented letter counts as one character. A table is filled in where cell (i, j) holds the cheapest way to turn the first i characters of A into the first j characters of B: d(i, j) = min(d(i−1, j) + 1, d(i, j−1) + 1, d(i−1, j−1) + c), with c = 0 when the two characters match and c = 1 when they do not. The corner cell is the distance d. Similarity is (1 − d / max(|A|, |B|)) × 100, shown as a percentage with one decimal. The comparison is case-sensitive and spaces count as characters. Two empty strings give no result, and so do pairs where the product of the two lengths exceeds 1,000,000 characters.

Common mistakes

• Expecting upper and lower case to match. Kitten and kitten are one edit apart because the capital K counts as a different character.
• Leaving a trailing space in a pasted string. A space is a character, so abc followed by a space against abc has a distance of 1.
• Reading similarity as a measure of meaning. It only compares characters in order, so two words that mean the same can score low.
• Treating the distance as the number of differing positions. Insertions and deletions shift everything after them, so the distance can be much smaller than a position-by-position comparison suggests.
• Pasting two long texts and expecting a figure. Once the two lengths multiplied pass one million, the tool gives no result, so compare shorter pieces.