JSON Diff

How it works

Both sides go through JSON.parse first, so if one isn't valid JSON you get the parser's own error message pointing at the problem instead of a garbage diff. The comparison then walks the two parsed structures recursively: for objects it takes the union of both key sets and recurses into each key, for arrays it compares index by index, and leaf values are considered equal only if they serialize identically. Every difference is emitted as a path like user.address[1].zip tagged as added, removed, or changed, then summarized into counts.

Because the diff runs on parsed values rather than text, formatting is invisible: indentation, key order, and whitespace differences produce zero noise, which is the whole point versus a line-based text diff. The trade-off to understand is array alignment. It's positional, so inserting one element at the front of a long array reports every following index as "changed" plus one addition at the end. When you see a wall of changed entries that all look shifted by one, read it as "something was inserted or removed early in this array", not as dozens of independent edits.

Parsing and comparison happen entirely in this page's JavaScript. Open DevTools' Network tab and hit Compare: no request carries your JSON, so API responses and config files with secrets in them stay on your machine.