JSON Analyzer — Size & Structure
Profile a JSON document: byte size, minified savings, key count, nesting depth, a type breakdown, and the largest fields — so you know what's costing bandwidth.
Know your payload before you ship it
Slow API responses and bloated config files usually have one or two culprits — a giant embedded string, a base64 blob nobody trimmed, an array with thousands of elements. The analyzer measures the document so you don't have to guess: the largest fields list ranks every value by its serialized byte size, so the thing costing you bandwidth is right at the top, with a click to jump to it in the source.
What each number means
- Formatted vs minified — read-size vs wire-size, and the whitespace that separates them.
- Keys / unique keys — total properties, and how many distinct names (a big gap hints at repetitive records).
- Max depth — how deeply nested the structure gets; very deep JSON is slow to traverse and awkward to model.
- Nodes & type breakdown — the overall value count and what it's made of, at a glance.
Then act on it
Found the heavy field? Minify for the quick win, extract just what you need with the JSONPath filter, or restructure. The analysis is a diagnosis; the rest of the toolset is the treatment.
Frequently asked questions
What does the analyzer measure?
Byte size (formatted and minified, with the whitespace savings), total and unique key counts, maximum nesting depth, the total node count, a breakdown of values by type, the longest array, and the largest individual fields by serialized size.
How is "largest field" calculated?
Each value is measured by the byte length of its JSON serialization, then ranked. This surfaces the parts of a payload actually costing bandwidth — usually a long string, a big embedded blob, or a large array — so you know what to trim first.
Formatted vs minified size — which matters?
Minified is what you send over the wire; formatted is what you read. The savings percentage is what minifying would remove. Note that if your server gzips responses (most do), the on-the-wire difference is smaller than the raw percentage suggests.
What counts as a node?
Every value: each object, array, string, number, boolean and null is one node. It is a quick proxy for how much structure — and therefore parsing work — a document carries.
Is my JSON uploaded?
No — the analysis runs entirely in your browser. Nothing is sent anywhere, so profiling confidential payloads is safe.