Diff checker

What a diff shows you

A diff is a precise, side-by-side answer to one question: what changed between two versions of a document? It's how code review, version control and document-comparison tools all work — instead of re-reading both copies and hoping to spot the edits, you get every removal and addition highlighted. This tool compares line by line (rather than character by character), which keeps the output readable and matches what you see in a pull request: removed lines in red, added lines in green, unchanged lines plain. It works equally well on source code, configuration files, prose drafts and exported data.

How line-based diffing works

Under the hood is the same idea git uses: a longest-common-subsequence algorithm. It finds the largest set of lines that appear in the same order in both texts — those are the lines that didn't change — and then everything left over on the original side is marked removed, and everything left over on the new side is marked added. Anchoring on the longest shared run is what stops a single early edit from making every line below it look different; it's why a good diff lines up the unchanged context around your actual changes. The trade-off is granularity: because it reasons about whole lines, changing one word inside a line shows that line as a removal plus an addition, not a word-level tweak.

Getting a clean diff

A line counts as unchanged only if it matches exactly, so trailing spaces, tabs-versus-spaces and different line endings can make identical-looking lines show as changed. The most common source of a noisy diff is reflowed text — when wrapping shifts, every wrapped line moves and the diff lights up. If the result looks messier than the real change, normalise both sides first: trim trailing whitespace, use one consistent indentation, and keep one sentence or statement per line so edits stay localised.

When a structural diff is better

Line diffing is the right tool when line order carries meaning — code, logs, plain text. It's the wrong tool when formatting is incidental: two JSON objects that are semantically identical but have keys in a different order, or different indentation, will show as completely different line by line. For that, compare by structure instead with the JSON diff tool, which ignores key ordering and whitespace. Either way the comparison runs entirely on your device, so it's safe for confidential files. When you need to reshape one side before comparing, the case converter and JSON formatter help normalise it first.