JSON field detection

Processed locally

app.inspect.view

You can also load data via #data={"name":"Ada"} or #url=... It is cleared from the address bar after it is read.

Scan JSON for values that look like timestamps, colors, or URLs and preview them. Logs and configs are full of these, and plain text is hard to judge at a glance. Unix second timestamps are especially easy to miss as ordinary integers.

Timestamps

Second- or millisecond-scale numbers can be shown as readable time. Handy for checking whether exp、createdAt looks reasonable.

Heuristic detection may treat ordinary large integers as times. Use the field name as well.

Colors / links

Detects common hex colors and URLs so you copy less between tools. Theme colors and callback URLs in config are easy to scan.

Common strings such as emails are flagged when possible.

Still local

Preview does not require uploading the whole JSON. Fits payloads that contain intranet URLs or login times.

Results are a list, not rewritten JSON. Keep editing the original on the formatter.

Example

JSON with special fields

{
  "lastLogin": 1720000000,
  "color": "#0052FF",
  "docs": "https://jsonsnap.com"
}

Detection results (illustrative)

lastLogin → 时间
color → 颜色
docs → 链接
Other fields display as ordinary strings / numbers.

What if detection is wrong?

Rules are heuristic. An ordinary business number may look like a timestamp. Product meaning wins.

How do I see hierarchy?

Use Tree view to collapse structure, then use this page for leaf fields.

Does it change the original JSON?

No. It lists matches and paths so you can review them.

What about ISO date strings?

Common ISO-8601 text is detected when possible. Unusual formats may be missed—search by field name on the tree.

Recommended workflow

Paste config / log JSON that contains times, links, or colors.
Review type, path, and preview value in the detection list.
Compare suspicious timestamps with the product time zone. Do not treat milliseconds as seconds.
To change values, go back to the formatter and edit the original.

This page answers reading questions like “is this number a time?” It is not a general data-labeling tool.