JSON Formatter & Validator

The formatter reads the whole document into memory and prints it again with two-space indentation. Because it rebuilds the document rather than re-indenting text, a few details can change:

• Numbers are written in their shortest form. 1.50 becomes 1.5, 1e3 becomes 1000, 1.0E-7 becomes 1e-7, and -0 becomes 0.
• Escapes are resolved. \u00e9 is shown as é and \/ as a plain /. Quotes, backslashes and control characters keep their escapes.
• Duplicate keys collapse silently. {"a":1,"a":2} comes out with only "a": 2, with no warning that the first value was dropped.

Numbers are read as standard double-precision values, so integers beyond 9007199254740991 are rounded:

The output looks valid, which makes this easy to miss. Order numbers and database keys of 16 digits or more are the usual victims. If you need to keep such values exactly, they should be sent as strings. Don't copy formatted output back into a system as the source of truth for those fields.

Keys normally stay in the order you wrote them, with one exception. Keys that look like whole numbers move to the front in ascending order:

The data is the same, since JSON objects have no defined order, but a diff against the original will show moved lines. Keep that in mind when you format two versions of a file to compare them.

Any single JSON value works, including a bare string or number such as 42. Two values one after the other do not, so JSON Lines files and log extracts with one object per line fail with an error. Format those one line at a time.

Error messages come from your browser's own JSON parser, so the wording differs between browsers. Most include a position or a line and column that points close to the problem. For the reverse job, stripping the whitespace back out, use the JSON Minifier.