Skip to content
TabBench

AI JSON Explainer

Analyze and explain JSON payloads, nested schemas, data structures, and potential security issues.

Runs on your device by default. The optional cloud AI mode sends your input to Google Gemini.

What the AI JSON Explainer does

Opening an unfamiliar API response or config file often means scrolling through hundreds of lines to work out what is in it. This explainer maps the whole structure in your browser: every field path with its type, the format of its values (dates, emails, URLs, IDs, IP addresses, tokens), what the field most likely means, and which fields are optional because only some items in a list have them. It also flags secrets and personal data, mixed types and numbers stored as text. Choose Cloud AI to add a written explanation from Google Gemini.

How to understand a JSON document

  1. Paste JSON or open a .json file. Use Format JSON to indent a minified payload.
  2. Leave the engine on On-device for a private structural analysis, or pick Cloud AI to add a written explanation.
  3. Press Explain JSON, or Ctrl + Enter. If the JSON is invalid, the exact line and column of the error are shown.
  4. Read the overview and the "Worth knowing" notes: secrets, personal data, inconsistent types and optional fields.
  5. Scan the field table, then copy or save the whole analysis as Markdown.

The AI JSON Explainer runs on your device by default. If you switch to the optional cloud AI mode, your input is sent to Google's Gemini model to produce the result.

When to use it

Integrating a new API

The field table shows at once that invoices[].amount is sometimes a number and sometimes a string, and that note is present in only some invoices. Those are exactly the details that cause bugs when you write code against the API.

Reviewing data before sharing it

Before pasting a payload into a ticket, a chat or an AI tool, the explainer flags fields like api_key and email so you can remove them first.

Documenting a payload

Saving the analysis as Markdown gives you a ready-made field reference for a README or wiki page, with types and examples filled in.

Good to know

  • Paste a real, complete sample. Fields that are missing or null in the sample cannot be described fully.
  • "[]" in a path means every item of a list, so orders[].total is the total of each order.
  • Field meanings are inferred from names and values. Treat them as a strong hint, and confirm with the API's documentation where it exists.
  • Cloud AI sends the JSON to Google. Leave it on On-device, or remove sensitive values first, for production data.

Frequently asked questions

Is my JSON validated using AI?

No, JSON syntax parsing is handled deterministically by the browser engine for 100% exact error detection, while AI provides the high-level explanation.

Is my JSON checked by an AI model?

No. Validation and the structural analysis use the browser's own JSON parser and run on your device, so the error positions and field list are exact. Only the optional written explanation comes from Google Gemini.

How does it decide a field is sensitive?

From the words in its name (password, secret, token, api_key, authorization and similar) and from its value, such as a signed JWT. Personal data is flagged from names like email, phone and address, and from values that look like emails, phone numbers or IP addresses.

How large a file can it handle?

Several megabytes comfortably. For very large arrays the first 500 items are analysed, which is plenty to capture the shape of the data, and the notes say when that limit was reached.

Related tools