Format raw JSON with indentation that is easier to scan, review, and share.
{
"name": "Softkey Tools",
"stack": [
"Angular",
"TypeScript",
"SSR"
],
"published": true
}Understand the format
Formatting rewrites only the whitespace between JSON tokens, so you can read a compressed payload without changing a single value.
JSON is defined by RFC 8259 as a token grammar in which whitespace between tokens is insignificant. Newlines, spaces, and tabs that sit between a comma and the next key can be added or removed freely without altering the document that a parser produces. That is the entire basis of pretty-printing: the formatter parses the text into values and then serialises those values again with indentation.
Because the round trip goes through a parser, formatting is also an implicit correctness check. If the input cannot be parsed, there is nothing to re-serialise, and the tool reports the parser error instead of guessing at a repair. A formatter never fixes broken JSON; it either reprints valid JSON or tells you why the text is not valid.
Indented JSON is for humans: code review, incident tickets, and documentation. Compact JSON is for transport and storage, where every byte of indentation is paid for on each request. The difference is smaller than it looks once HTTP compression is applied, because gzip and brotli collapse repeated runs of spaces very effectively, but it still matters for large payloads, message queues, and log lines that must stay on one line.
One practical rule: store and transmit compact, review and diff pretty. If a fixture file is committed to a repository, indented output produces far more readable version-control diffs, so the reviewability argument usually wins there.
Step by step
Formatting runs entirely in your browser using the built-in JSON parser. The payload is never transmitted, logged, or stored.
Worked examples
Input
{"id":1042,"paid":true,"items":[{"sku":"A-1","qty":2}],"note":null}Result
{
"id": 1042,
"paid": true,
"items": [
{
"sku": "A-1",
"qty": 2
}
],
"note": null
}Nothing was added or removed. The nesting is now visible, and it is obvious that "paid" is a boolean while "note" is an explicit null.
Input
{
"query": "status:open",
"limit": 25
}Result
{"query":"status:open","limit":25}Compact mode is useful when a payload must travel inside a single log line, a shell argument, or an environment variable.
Reference
| Type | Example | What to watch for |
|---|---|---|
| string | "order-1042" | Double quotes only. Control characters, quotes, and backslashes must be escaped. |
| number | 42, -3.5, 1.2e6 | No NaN, Infinity, hex, or leading zeros. Values beyond 2^53 lose precision in JavaScript. |
| boolean | true, false | Lowercase only. "true" in quotes is a string, not a boolean. |
| null | null | Distinct from an absent key and from an empty string. |
| object | { "id": 1 } | Keys must be quoted strings. Key order carries no meaning. |
| array | [1, 2, 3] | Ordered. Mixed types are legal but usually a schema smell. |
Practical Guide
Troubleshooting
FAQ
No. Only insignificant whitespace changes. Keys, values, types, and array order are preserved exactly, because the output is produced by re-serialising the parsed document.
Object key order is preserved as it appeared in the input, but two systems may emit keys in different orders. Compare parsed structures rather than raw text when that happens.
No. Comments and trailing commas belong to JSON5 and JSONC, which are separate formats. Strict JSON parsers reject both, so they must be removed first.
It is a convention, not a rule. Two spaces keeps deeply nested payloads narrow enough to read; four spaces is common in language ecosystems that already use it. The parsed result is identical either way.
No. Parsing and re-serialising both happen locally in the page, so nothing is uploaded to a server.
Go deeper