JSON comparator

Compare two JSON documents after normalizing layout and object property order. A one-line API response may therefore match the same response formatted differently. Value types and array order still matter: "1" is not the number 1, and a reordered list is different. Both sources stay in the browser. Output is a readable text report, not a new JSON document or a patch to apply automatically.

Both inputs must follow JSON syntax. The engine parses values, recursively sorts object keys and serializes them with common formatting. Layout whitespace and written property order stop producing artificial differences. Whitespace inside strings remains significant.

Arrays are not sorted and their objects are not matched by identifier. Moving an entry remains a difference. Strings, numbers, booleans and null are compared as parsed values. A missing key differs from a key containing null. Duplicate names follow JSON parsing behavior, retaining the last occurrence; avoid that ambiguous representation.

When normalized forms differ, each line is compared with the same position in the other document. - identifies the first input and + the second; shared lines have a space prefix. An insertion can shift several lines, producing a longer report than the smallest logical change. This is neither a path-based semantic diff nor JSON Patch.

Unsafe JavaScript integers and numbers parsed as nonfinite are rejected to avoid declaring distinct rounded identifiers equivalent. Decimal values still follow browser numeric precision. Use contractual string representations for long identifiers; choose text comparison when exact written tokens matter.