Paste two versions of a text: a config before and after an edit, two drafts of a contract, a SQL query from two branches. Lines are matched with the Myers algorithm (the one git diff uses by default), and for every changed line the exact words or characters that differ are highlighted.
Unchanged stretches longer than a few lines are folded so the changes stay visible. With Ignore whitespace changes on, reindented code and trailing spaces no longer count as differences.
How to use
- Paste the original text in the left field and the changed one in the right field.
- Pick side-by-side or unified view. On a phone unified is easier to read.
- Turn on ignoring whitespace or case if those changes do not matter to you.
- Click a folded block to expand it, or copy the whole diff with Copy.
Examples
apple banana cherry vs apple blueberry cherry date → +1 added · −0 removed · ~1 changed
"Total: 100" vs "total: 100" → No differences
Frequently asked questions
What counts as a changed line versus added and removed?
When a block of removed lines is directly followed by added lines, they are paired in order and shown as changed, with an inline diff. Lines left over without a pair count as added or removed.
How are differences inside a line found?
Short lines (up to 32 characters) are compared character by character, longer ones word by word. Punctuation and spaces count as separate tokens, so a moved comma does not mark the whole word.
Does a missing newline at the end matter?
No. One trailing newline is ignored, and Windows (CRLF) and Unix (LF) line endings are treated the same.
How large can the texts be?
Several thousand lines work fine. The result is recomputed after you stop typing, so the fields do not lag. Two completely different 5000-line texts take a fraction of a second.