Text Diff Checker
What changed between two versions — added, removed, unchanged.
How the comparison works
Both inputs are split into lines and compared with a longest-common-subsequence algorithm, which finds the largest set of lines that appear in both versions in the same order. Everything not in that set is either an addition or a removal. The result is a classic unified diff: lines prefixed with a minus exist only in the original, lines with a plus only in the revised, and unprefixed lines are unchanged. This tool works line by line, so a change inside a single long line shows as one removal plus one addition rather than highlighting the specific words.
Why line-based diffing has limits
Reordering a file produces a large diff even though the content is identical, because moving a block destroys the common subsequence in that region. Reformatting — wrapping paragraphs differently, changing indentation, converting tabs to spaces — creates noise that hides the substantive change underneath. This is why most teams normalise formatting before diffing, adopt a shared formatter, and configure their diff tool to ignore whitespace-only changes. If your diff is enormous after a change you believe was small, suspect formatting first.
Common uses
- Reviewing edits. What did the AI actually change in my draft? Paste the original and the rewritten version.
- Comparing configurations. Two versions of a config file, a YAML deployment manifest, or an environment file.
- Verifying data. Two exports from different dates, to see which rows changed.
- Checking translations. Which strings were modified between releases.
- Debugging output. The same prompt run twice against different model versions.
Reading a diff well
The temptation is to skim for surface changes, but the important part is usually what is absent. A line removed from a configuration file may disable a check that nothing else references. A changed default may affect every caller. When reviewing, ask what each removal permitted rather than only what each addition does. On substantial changes, diff the smaller logical units separately — a single coherent change is far easier to review correctly than five interleaved ones.
Privacy and size
Everything happens in your browser tab; neither version is transmitted anywhere, so this is safe for contracts, credentials and unreleased code. Because the comparison runs entirely in the page, very large inputs will be slow — for files above a few thousand lines, a local diff tool or Git will handle it better. Before comparing, it often helps to normalise with the line cleaner or to strip variable parts using the regex tester.