JSON Diff (Semantic Compare)

RFC 8259 defines a JSON object as an unordered collection of name/value pairs, and insignificant whitespace can appear anywhere. Two API responses can therefore be byte-for-byte different and still carry identical data: one serializer writes {"id":1,"name":"a"}, another writes the keys in the opposite order with two-space indentation. A line-based diff, like the Diff Checker on this site, reports every one of those lines as changed.

This tool parses both sides first and compares the resulting values. Reordered keys and reformatting produce no differences at all, and what remains is the data that actually changed. That makes it the right choice for comparing API responses between environments, reviewing config file changes, checking a migration's output against a snapshot, or verifying that a refactor of a serializer did not change its output.

Each difference is listed with a JSONPath such as $.dimensions.w or $.items[2].sku. Keys that are not valid identifiers use bracket notation, for example $["content-type"].

• Added: the path exists only on the right.
• Removed: the path exists only on the left.
• Changed: same type, different value, such as 19.99 to 24.99.
• Type changed: the JSON type differs, such as the number 19.99 becoming the string "24.99", or null becoming 0. These are often the bugs that break typed clients, so they are called out separately.

A key set to null is not the same as a missing key; the diff reports the difference because many APIs treat them differently.

Arrays are compared by position by default, which is correct for ordered lists such as steps or coordinates. Choose Ignore order for sets such as tags or permission lists; elements are then matched by value (duplicates are counted), and only elements with no match are reported.