Text Diff

A text diff lays two versions of a document against each other and marks exactly which lines were added, removed or left alone, so one changed line cannot hide in an identical page. Paste the original left and the revision right, press compare, and each line gets a verdict: green for added, red for removed, neutral for unchanged context. Underneath, a longest-common-subsequence table over both line sequences is walked into the alignment, so scattered edits and moved blocks surface as precise line operations, not a vague score. A totals line reports added, removed and unchanged counts, and identical inputs honestly report no difference.

Reading the alignment output

• ➕ Added lines in green, present only in the revised version
• ➖ Removed lines in red, present only in the original
• ✔️ Unchanged lines kept as context around every change
• 📊 Running totals summarizing the size of the edit
• ⌨️ Explicit compare button, so big pastes never work mid-typing

Reviews where it proves its worth

• Seeing what a reviewer actually altered in your draft
• Auditing a config file before and after a find-and-replace sweep
• Confirming two extracts differ only where expected

Note the comparison is line-based: a one-word edit flags its whole line, and moved lines read as removed plus added rather than moved. The whole diff runs in your browser with neither version uploaded, safe for contracts, configs and unpublished manuscripts alike. Identical inputs report no difference instead of a misleading empty screen, so a clean result reads as a verdict.

• A longest-common-subsequence over lines, grouped into change blocks — the classic diff model, computed locally.
• Line mode never matches partial lines: one changed word marks the whole line, keeping reviews honest.
• Whitespace-only churn is flagged separately so indentation noise cannot hide real edits.