JSON Minifier

What this tool does

Minifying JSON means removing every byte that only exists for human readers: the newlines between members, the indentation in front of them, and the space after each colon and comma. Nothing else changes. The parsed value that comes out is identical to the one that went in.

The tool shows what that is worth — original size, minified size, the percentage saved and, where your browser supports it, the gzipped size of the result — and can additionally emit the JSON as an escaped string literal or percent-encoded, which is what you need when the document has to survive a config file or a URL.

Common uses

• Packing a JSON value into an environment variable, a Kubernetes secret or a CI variable, where a multi-line value is awkward or forbidden.
• Shrinking a fixture that is committed to a repository and read by machines only.
• Putting a small JSON payload into a query string or a data attribute, where the escaped and URL-encoded forms matter more than the raw one.
• Trimming a response body before pasting it into a ticket or a chat message.
• Measuring the real cost of pretty-printing a payload your API returns on every request.

A short example

72 bytes of formatted JSON:

{
  "id": 7,
  "tags": [
    "alpha",
    "beta"
  ],
  "active": true
}

becomes 46 bytes — a 36% saving:

{"id":7,"tags":["alpha","beta"],"active":true}

What minifying does and does not change

The output is produced by parsing the document and re-serialising it, which has two visible consequences beyond whitespace removal:

• Numbers are re-serialised. 1.0 becomes 1, 1e3 becomes 1000 and 1.50 becomes 1.5. The values are equal; the text is not. If you are comparing checksums of the document rather than of the data, that difference matters.
• Escapes are normalised. A string written as "caf\u00e9" comes back as "café", because only ", \ and the control characters below U+0020 have to be escaped. Both spellings are correct JSON for the same string, but a byte-for-byte comparison will disagree.

Key order is never touched, so a minified document diffs cleanly against another minified document from the same source.

Worth knowing

Over HTTP, minifying is worth much less than it looks. Whitespace is the most compressible thing in a document — long runs of identical spaces are exactly what DEFLATE encodes in a few bits — so a payload that shrinks 30% when minified often shrinks only 2–5% more after gzip or Brotli compression on top. If your API already sends Content-Encoding: gzip, spend the effort on the number of fields you return instead. Minifying still wins where compression is not in play: env vars, URLs, database columns and log lines.