JSON Formatter Online Free

A JSON formatter with its own hand-rolled syntax highlighter and a repair pass for the mistakes people actually make

Formatting valid JSON is one line: JSON.stringify(parsed, null, indent). This tool goes further. It uses its own strict parser, so large numbers like 12345678901234567890, values like 1.50 and duplicate keys come out exactly as you wrote them, which JSON.stringify cannot promise. It also has to color every token for the highlighted view, render a collapsible tree, count keys and nesting depth, and attempt repairs on JSON that almost, but does not quite, parse. Each of those is a distinct piece of logic worth understanding.

Everything runs in your browser. There is no network call anywhere in the script, so pasted JSON, however sensitive, never leaves the page.

The highlighter is a hand-written character scanner

Rather than using a regex to color tokens after the fact, the highlighter walks the already formatted JSON string character by character and classifies each token as it goes: strings, then a lookahead check for a following colon to decide if a string is an object key or a value, booleans and null matched against fixed length substrings, numbers matched by scanning digits, a decimal point, sign, and exponent marker, and finally structural punctuation.

// simplified from the actual highlighter, key vs value string detection
var j = i; // after closing quote
while (j < len && ' \n\r\t'.indexOf(json[j]) >= 0) j++;
if (json[j] === ':') {
  res += '<span class="key">' + str + '</span>';
} else {
  res += '<span class="value">' + str + '</span>';
}

Three views, one parsed object

Formatting parses the input once. The resulting tree is kept in memory and reused across all three output modes, so switching views does not reparse anything.

The repair pass, and what it honestly cannot fix

Clicking repair first tries a strict parse. If that succeeds, nothing else happens. If it fails, a lenient tokenizer reads the input the way a JavaScript engine would and writes it back as strict JSON: comments are dropped, single-quoted strings become double-quoted, unquoted keys get quotes, trailing commas go, missing commas between items are added, Python style True, False and None become true, false and null, and brackets left open at the end are closed. Because it works token by token, an apostrophe inside a string or a URL like http://example.com is never damaged.

Extra actions beyond formatting

The stats row reports total object keys, total arrays, maximum nesting depth, and byte size, all computed by one recursive walk that tracks the deepest depth reached and increments counters for every key or array it passes through.

JSON specifications and tooling

• json.org is the canonical grammar diagram for JSON, useful for seeing exactly what counts as a valid number, string, or structural token.
• RFC 8259 is the current IETF standard defining the JSON data interchange format.
• MDN, JSON.stringify documents the indent argument this tool uses for pretty-printing.
• JSON5 is a real extended grammar supporting comments, trailing commas, and unquoted keys natively, worth reaching for if you need to parse this kind of relaxed JSON regularly rather than repair it after the fact.

Everyday JSON tasks

Pretty printing an API response copied from browser devtools, cleaning up minified config files before editing them, fixing JSON that was hand typed with single quotes or trailing commas, comparing structure with the tree view before writing a schema, and checking payload size before committing to a request body limit.