Longest common subsequence
Text is split into words (and CJK characters) and compared with an LCS matrix — the same idea code review tools use.
$ compare | find | differences
Compare two pieces of text.
Comparing…
Text is split into words (and CJK characters) and compared with an LCS matrix — the same idea code review tools use.
Very large comparisons are throttled and fall back to line-level diff with a clear on-page notice instead of freezing your browser.
Send the revised text to Word Counter in one click. The hand-off uses a one-time session value that is deleted as soon as it is read.
Reading two drafts side by side and hunting for changed words is slow and error-prone, especially in review comments where you need a short, precise description of the edits.
The default inline view reads like one continuous document. Switch to side-by-side or turn on code mode when you are comparing long documents or source files.
Original: The quick brown fox jumped over the lazy dog near the river. Revised: The quick red fox leaped over the sleepy dog near the old river. => 4 changes: brown → red, jumped → leaped, lazy → sleepy, and the added word “old”.
Adjacent deletions and insertions are grouped into one change block, so a replacement such as brown → red counts once.
Word splitting makes source code hard to read. Code mode compares whole lines instead, which matches how developers review code.
Word-level comparison is intended for texts under about 5,000 words. Above that the page suggests comparing in smaller sections.
No. The comparison runs locally in JavaScript and neither version leaves the browser.