utilyfe

JSON Formatter

Pretty-print JSON, unescape it from a log line, or find out exactly where it breaks.

Runs entirely in your browser. Nothing you enter is uploaded or stored.

Also try:Base64 Encoder and DecoderURL Encoder and DecoderUUID Generator

How to use the JSON Formatter

  1. 1Paste your JSON into the box, or press "Try an example" to see it work.
  2. 2Pick a mode: Format to tidy and check it, Unescape for JSON copied out of a log or a string field, Escape to turn a document into a string you can embed.
  3. 3If the JSON is broken, the line and column of the problem appear under the box, with a note on the likely cause.
  4. 4Choose your indent (2 spaces, 4 spaces or a tab) or tick Minify to put everything on one line.
  5. 5Press Copy to take the result. Your original text stays in the top box.

What counts as valid JSON

JSON is a much stricter format than the JavaScript it came from. The rules that catch people out:

  • Double quotes only. Both keys and text values need them. 'name' and name are both rejected; only "name" is valid.
  • No trailing commas. The comma in [1, 2, ] is an error, even though your editor may accept it elsewhere.
  • No comments. There’s no way to write one in standard JSON.
  • Numbers are plain. No NaN, no Infinity, no leading +, no hex, no trailing dot. 01 is invalid too.
  • undefined doesn’t exist. Use null.
  • Any value can sit at the top level. An object is usual, but a bare array, string, number, true, false or null is also valid JSON.

Errors and what they usually mean

What you’ll see What’s usually wrong
Expected a key in double quotes, with a trailing-comma note A comma sits before the closing }.
Expected a key in double quotes: single quotes and bare words aren’t valid 'name' or name where "name" is needed.
Two values need a comma between them, or a bracket is left open A missing comma between items, or a copy-paste that stopped short.
The JSON ends too early The text runs out mid-value: a brace, bracket or quote is never closed.
There’s extra text after the JSON value ends A second object pasted after the first, or a stray character at the end.

The line and column we show point at the character the parser choked on, which is where it gave up rather than where you slipped.

Escaped JSON, and how to get rid of it

JSON has no problem carrying JSON inside itself, but only as text, with every quote escaped. It turns up constantly:

  • a log line that records a request body as one string
  • a webhook whose body or message field holds the real payload
  • a queue message, where the useful part is Message inside the envelope
  • a database column that stores JSON as a string

The result is unreadable, and the backslashes multiply every time it’s wrapped again:

Level What you see
The data {"orderId":123}
Escaped once "{\"orderId\":123}"
Escaped twice "\"{\\\"orderId\\\":123}\""

Unescape peels those layers off, up to five of them, and tells you how many it removed. It works whether or not the value still has its surrounding quotes, so a fragment copied out of the middle of a log line is fine.

Expand JSON inside strings is for the other shape: a normal document where only some fields hold escaped JSON. Tick it in Format mode, and each of those fields is parsed in place, so the whole payload reads as one document.

Escape does the reverse, turning a document into a single string with the quotes and line breaks escaped, ready to paste into another payload or a test fixture.

Sorting keys makes two payloads comparable

Tick Sort keys A–Z and every object is rewritten in alphabetical order, at every level. This is the quickest way to compare two API responses that contain the same data in a different order: sort both, then diff them. It changes only the order of keys, never their values.

Large numbers lose precision

JSON numbers are parsed as standard floating-point values, which hold about 15–17 significant digits. A 20-digit ID, like a Twitter or Discord snowflake, comes back slightly wrong, and formatting writes the wrong value back out. This isn’t a quirk of this page: JSON.parse behaves the same way in Node and in every browser.

If your data has IDs that long, they should travel as strings ("1234567890123456789"), which survive untouched.

Minify or pretty-print

Pretty-printing is for reading and for diffs: one value per line, indented by nesting. Tabs suit teams who want everyone to control their own indent width; 2 spaces is the common default.

Minifying strips every space and line break. It’s for sending data over the network or pasting into an environment variable, not for storing in a repository, where a one-line file makes code review painful.

Questions

Is my JSON uploaded anywhere?
No. The parsing and formatting happen in this browser tab, which matters when the JSON is an API response or a config file with keys in it. Nothing is uploaded, logged or saved, and refreshing the page clears the box.
Why does the error point at a line that looks fine?
A JSON parser only notices a problem when it reaches something it can't accept, which is often one line after the real mistake. A trailing comma is the classic case: the comma on line 8 is only reported when the closing brace arrives on line 9. Check the line above the one we name.
My API returns JSON inside a string, full of backslashes. Can this read it?
Yes, two ways. If the whole thing is escaped (a value copied out of a log line, say), switch to Unescape and it peels the backslashes off, including when it has been escaped twice. If the escaped JSON sits in one field of an otherwise normal document, keep Format and tick "Expand JSON inside strings", which parses those fields in place so you can read the whole payload at once.
Does it support comments, trailing commas or JSON5?
No, and that's deliberate. This validates against the JSON standard, which is what your API or config loader will use. If your file has comments, it's JSONC or JSON5. Strip them before pasting, or the parser will stop at the first slash.
Does formatting change my data?
Key order is kept unless you tick "Sort keys A–Z", and text, booleans and null pass through untouched. The one exception is very large numbers, which lose precision. That's explained above.
How large a file can it handle?
Files of a few megabytes work. Above that the tab will feel sluggish while it reformats on every keystroke, because everything runs on your own machine rather than a server.