JSON Diff Visualizer

Unsupported input

The tool may reject input that does not match the expected content, structure, or file type.

Fix: Confirm the tool input requirements and paste the correct type of data.

Incomplete values

Missing fields or partial content can block processing or produce weak results.

Fix: Provide the full required input before running the tool.

Copying placeholder content

Sample or placeholder values can lead to output that looks valid but is not ready for real use.

Fix: Replace placeholders with your actual values before relying on the result.

How is this different from the existing Diff Checker?

Diff Checker is a line-by-line text diff (Myers / Patience) — fine for code, but it produces noisy output for JSON because reformatted whitespace and reordered keys all show as changes. JSON Diff Visualizer parses both sides and walks the resulting object tree, so the diff is on the data itself: a moved key inside an object is a 0-cost change, not five removed lines plus five added lines.

What does 'ignore array order' actually do?

Each array item is hashed by its `id`, `_id`, `uuid`, `key`, or `name` field if present, otherwise by primitive value. Items with the same hash are matched even if their indices differ — and the result includes a 'moved' edge that records the new index. Without this toggle, JSON Diff falls back to positional comparison.

Can I generate a patch I can apply with `fast-json-patch`?

Yes. The 'JSON Patch' tab outputs a strict RFC 6902 array of operations (`add`, `remove`, `replace`, `move`). You can paste that into any RFC 6902 implementation — `fast-json-patch`, `rfc6902`, JSON Patch in `jsonpatch-py`, or as an HTTP `PATCH` body.

Will it diff two huge JSON documents?

Comfortably up to a few MB per side. Both editors are Monaco with virtualization. The diff itself is O(n) over the data plus O(n log n) for array hashing; tens of thousands of nodes are usually fine, but pathological inputs (deeply nested + many moves) get slower.

Are my JSON documents sent to a server?

No. Parsing, diffing, and patch generation all happen in the browser — there is no server round-trip. The same applies to file uploads; the file is read with `FileReader` and never leaves your device.

Why are some 'unchanged' nodes still shown?

Unchanged nodes are hidden by default unless they're an ancestor of a change. Toggle 'Unchanged' in the toolbar to expose every key — useful when you want to view the structure of both documents side-by-side, not just the deltas.