JSON Diff Checker

JSON Diff Checker compares two JSON documents structurally rather than line-by-line. Instead of telling you "line 14 differs", it parses both documents and reports exactly which keys were added, removed, or changed - with full paths like features.rateLimit or tags[2] - so you can see at a glance what changed between two API responses, config files, or package manifests.

Text-based diff tools produce noise on JSON: reordered keys or different indentation show up as changes even when the data is identical. A structural diff ignores formatting entirely and compares the parsed values. It also catches a class of bugs plain diffs hide, like a value silently changing type from the number 3000 to the string "3000" - flagged here as a distinct "type changed" entry, because that difference breaks strict-typed consumers even though it looks identical at a glance.

The comparison runs client-side in your browser. The JSON you paste - API payloads, environment configs, database exports - never leaves your device, so it is safe to compare documents from internal or production systems. Clicking Compare sends only a usage check and a record of how many differences were found. There is no signup, no upload, and the tool works offline once the page is loaded.

Load Sample shows what the JSON diff checker reports when you compare two JSON files that differ in every way. The original has "version": "1.4.0", "port": 3000 and rateLimit 100; the updated file has "1.5.0", "port": "3000", rateLimit 250, a new metrics key and a third tag. Compare returns 5 differences: +2 added (features.metrics, tags[2]), ~2 changed (version, features.rateLimit) and !1 type changed (port). Copy Report turns that into five marked lines, starting with ~ version: "1.4.0" → "1.5.0".

Comparing structurally rather than as text is what makes the result readable, and it rests on a fact about JSON that a text diff cannot use: object keys are unordered by specification, so two documents that differ only in key order are the same document, and a line-based diff will report every line as changed. Arrays are the opposite, since order in an array is meaningful, so a reordered list is a real difference by default. The option to ignore array order exists for the common case where an API returns a set in whatever sequence it likes. Numbers are compared by value, so 1.0 and 1 match, which is usually what you want and occasionally hides a type change. Type changes are reported separately for that reason, since a field moving from the number 5 to the string "5" breaks consumers while looking almost identical.