Find what actually changed.

Compare structured data. Match records, skip noise, and inspect every difference.

Local processing · 5 MiB per document

Before

After

Match records by field

One top-level field per record, applied to all record arrays. IDs must be unique strings or numbers. Primitive arrays use position.

Ignore field names

Comma-separated, case-sensitive names at every depth. Object key order is ignored. CSV values remain strings.

Ready when you are. Paste two documents or load the example.