If I just reorder keys without changing any values, will that show as a difference?

Yes. The comparison works on the two documents as formatted text, line by line, not as a structural match of keys regardless of position, so a key that moved shows as one removed line and one added line even though its value never changed.

How does the tool know which side has the error when only one document is invalid?

Each side is parsed on its own rather than in one combined attempt, so a syntax error is attributed to the exact editor, left or right, where it actually is.

JSON Diff Tool: Compare & Highlight JSON Differences Online

Originalpaste or type · or drop a .json file

Browse

Modifiedpaste or type · or drop a .json file

Browse

+2 additions-1 deletion

 {
   "id": 1,
   "name": "Alice Martin",
-  "role": "admin",
+  "role": "editor",
   "active": true,
+  "age": 30
 }

Why compare JSON rather than raw text

A plain text diff treats {"b":2,"a":1} and {"a":1,"b":2} as different, but they're the same JSON object. A JSON-aware diff normalises key order and data types before comparing.

You see only semantic differences: a field that changed value, a key that was added, an array element that was removed — no noise from formatting or key reordering.

Reading the diff output

The +/− counters above the diff view tell you at a glance how many additions and removals there are. Paste the two JSON documents and the diff updates immediately.

RemovedError tone — appears only in the left document

AddedValid tone — appears only in the right document

UnchangedNeutral color

Debugging and review use cases

Developers compare API responses across environments to catch configuration drift between staging and production. QA engineers diff before/after snapshots of database records to verify a migration.

Backend teams review the delta between two versions of a config file before deployment. Because both documents stay in your browser, you can paste sensitive payloads without concern.

Frequently asked questions

No. Both sides are re-serialized with the same 2-space indentation before comparing, so only actual differences in the data show up, not how either document happened to be formatted originally.