JSON formatter & validator

Where it runs

In your browser, using the same JSON.parse your own code uses. Nothing is uploaded and nothing is stored: load the page, disconnect the network, and it still works. That matters for this tool more than most, because the JSON a developer needs to read is usually a real response from a real system, and quite often has somebody’s data in it.

Using the browser’s own parser is a deliberate choice rather than a shortcut. A hand-written parser would be a second opinion about what counts as valid JSON, and the opinion that matters is the one held by the runtime your code will use.

The four things that are not JSON

Almost every “but it looks fine” turns out to be one of these. All four are perfectly good JavaScript, which is exactly why they catch people out.

• Trailing commas. {"a": 1,} and [1, 2, ] are both rejected.
• Single quotes. JSON strings and keys use double quotes, always.
• Unquoted keys. {a: 1} is an object literal, not JSON.
• Comments. There is no comment syntax in JSON at all.

Two things JSON quietly changes

Duplicate keys collapse. {"a":1,"a":2} is valid, and every parser keeps the last one. The formatted output has a single key and the first value is simply gone. The tool does not do this to you. Your code does the same thing, but it is worth seeing before you assume the output matches the input.

Large numbers lose precision. JSON has no integer type, so numbers become double-precision floats holding about fifteen significant digits. An ID of 12345678901234567890 comes back as 12345678901234567000. Anything above 9,007,199,254,740,991 has to be quoted as a string in the JSON itself to survive; no formatter can recover one that was not.

Sorting keys, and when to

Key order carries no meaning in JSON, so sorting changes nothing about what a document says. What it changes is comparability: two config files written by different people, or two API responses from different versions, become byte-identical if their contents match, and a diff shows only real differences instead of reshuffling.

Array order is never touched, because there it is the data. An array is a sequence and reordering it would be changing the document rather than tidying it.