Which JSON Schema version is generated?

Draft-07. The output opens with "$schema": "http://json-schema.org/draft-07/schema#" and uses the classic type / properties / required vocabulary, so Ajv and most validators accept it as-is.

How are required fields decided?

Every key present in the sample object is listed as required. Inference runs on one document, so a key that happens to be missing from your sample will not be marked required — edit the schema if your API may omit it.

How are nested objects and arrays typed?

Each nested object becomes its own schema block, and arrays become an "array" with an items schema inferred from their elements.

No. Inference and schema generation run in the page; the document never leaves the browser.

JSON Tool 02

JSON to JSON Schema

Turn a sample document into a JSON Schema with inferred types, required keys and nested structure — then copy or download it.

Input Ready

Sample JSON in, schema out

or paste your own JSON

Schema Waiting for input

Schema appears here

Types, required fields and nesting are inferred as you convert

Types, required keys and nesting are inferred from your sample document. Client-side only

About JSON Schema conversion

JSON Schema is a vocabulary for describing and validating JSON documents. This tool infers the structure of your sample — types, arrays, nested objects — and writes a draft-07 schema you can copy or download.

Example input

{
  "name": "John Doe",
  "age": 30,
  "active": true,
  "hobbies": ["reading", "coding"],
  "address": {
    "street": "123 Main St",
    "city": "New York"
  }
}

Generated schema

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "name": { "type": "string" },
    "age": { "type": "number" },
    "active": { "type": "boolean" },
    "hobbies": {
      "type": "array",
      "items": { "type": "string" }
    }
  }
}

Frequently asked

Which JSON Schema version is generated?: Draft-07. The output opens with "$schema": "http://json-schema.org/draft-07/schema#" and uses the classic type / properties / required vocabulary, so Ajv and most validators accept it as-is.
How are required fields decided?: Every key present in the sample object is listed as required. Inference runs on one document, so a key that happens to be missing from your sample will not be marked required — edit the schema if your API may omit it.
How are nested objects and arrays typed?: Each nested object becomes its own schema block, and arrays become an "array" with an items schema inferred from their elements.
Is my JSON uploaded?: No. Inference and schema generation run in the page; the document never leaves the browser.