JSON structure analyzer
Report a JSON document's structure — nesting depth, node counts, key frequency, per-path types, and array stats. Runs in your browser, no upload.
About this tool
JSON structure analyzer parses a JSON document and reports its shape — without transforming or reformatting the data. Paste any JSON value (object, array, or scalar) and you get:
- Max nesting depth — the deepest level of objects/arrays.
- Node counts by type: objects, arrays, strings, numbers, booleans, and nulls, plus total values, total keys, unique keys, and empty values.
- Key frequency — how often each key name occurs across the whole document (recurring names counted each time), ranked most-common first.
- Per-path type distribution — each path (with array indices collapsed to
[]so every element of an array shares one path) and the JSON types seen there. This makes mixed-type fields obvious. - Array stats — count, min / max / average length, and total elements.
- Byte size — raw vs. minified, plus the compression potential of stripping whitespace.
- Quality warnings — deep nesting (more than 5 levels), recurring keys, and empty values.
Worked example
For {"users":[{"id":1,"name":"Ada"},{"id":2,"name":"Bo"}]} the report shows root type
object, max depth 3 ($ → $.users → $.users[] → $.users[].name), key
frequency id: 2, name: 2, users: 1, and the path $.users[] with 2 objects. The
name field lands at $.users[].name as a string seen twice.
Options
- Output format — JSON report (structured, easy to diff or feed to another tool) or Plain text (a human-readable summary of the same data).
- Max keys and max paths — cap the key-frequency and per-path lists (set
0to list everything). A truncation flag tells you when a list was cut.
Privacy
Everything runs in your browser via WebAssembly — your JSON is never uploaded. The tool is also available from the gizza CLI and in chat, which return the report as structured JSON.
FAQ
How is nesting depth counted?
The root value is depth 0, and each step into an object or array adds one. So
{"a":1} is depth 1, {"a":{"b":1}} is depth 2, and a scalar like 42 on its
own is depth 0. Depth over 5 levels raises a "deep nesting" warning, since deeply
nested data is harder to read and query.
What does the per-path type distribution show, and why collapse array indices?
Each path is written with $ for the root, .key to descend an object, and [] to
descend an array — so all elements of an array share one path ($.users[] rather
than $.users[0], $.users[1], …). That collapse is what makes a mixed-type field
jump out: if $.items[].price reports ["number", "string"], some prices are strings.
Paths are ranked by how many nodes were seen at each one.
What counts as an "empty value"?
An empty value is a null, an empty string (""), or an empty object/array ({} /
[]). They're counted together in the empty_values total and surface a warning, which
is handy for spotting placeholder fields or optional data that never got filled in.
Can it handle very large JSON, and what about invalid JSON?
Large documents are fine — parsing and the single walk are linear, and the whole thing runs locally so there's no upload size limit beyond your browser's memory. If the input isn't valid JSON, you get a clear parse error (with the position) instead of a partial report, so you know exactly what to fix.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool json-structure-analyzer '{ "users": [ { "id": 1, "name": "Ada" }, { "id": 2, "name": "Bo" } ] }'New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/json-structure-analyzer/?json=%7B%20%22users%22%3A%20%5B%20%7B%20%22id%22%3A%201%2C%20%22name%22%3A%20%22Ada%22%20%7D%2C%20%7B%20%22id%22%3A%202%2C%20%22name%22%3A%20%22Bo%22%20%7D%20%5D%20%7D&format=json&top_keys=30&top_paths=50Machine-readable descriptor: tool.json — title + parameters JSON Schema for agents.