Measure how different two texts are

"Almost the same" is a number — Levenshtein distance counts the edits between two strings, and the percentage makes it readable.

Tools used in this guide

The job

Did the user type the same name twice? Is this error message a variant of a known one? Edit distance quantifies similarity where diff tools show too much detail.

Compare

Paste text A and text B. The tool reports the edit distance (minimum single-character insert/delete/substitute count) and a normalized similarity percentage.

"kitten" → "sitting" is distance 3 — the textbook example.
Similarity = 1 − distance ÷ max length; identical strings read 100%.

Worked example

"Smith, John" vs "Jon Smith" reads distance 10, similarity 9% — same person, different format; a threshold match would need fuzzier logic than raw distance.

Where it stops

Each text is compared on up to 3,000 characters (over-length input is flagged). Case and whitespace count as edits — normalize first if you want case-insensitive comparison. Word-level and semantic similarity are out of scope.