What version of JSON Schema does this generate?

You can choose between Draft 7 and Draft 2020-12. Draft 7 is the most widely compatible and is used by OpenAI structured output, Anthropic tool use, and OpenAPI 3.0. Draft 2020-12 is the latest stable release, used by OpenAPI 3.1 and modern validators. The schema structure is identical for both; only the $schema URI changes.

How does it handle arrays with multiple objects?

When an array contains more than one object, the tool merges all the item schemas together. Fields that appear in every object become required. Fields that appear in only some objects are included in properties but left out of required.

What does Detect formats do?

When this option is on, the tool checks string values against known patterns. If a value looks like an email address, a URL, a date, or a datetime, it adds a format keyword to that field. This is useful when the schema feeds a typed validation pipeline or structured AI extraction.

Can I use the output with the Spidra API?

Yes. The Spidra scraping and batch APIs accept a schema parameter. Paste the generated schema into your request and the AI extracts data in exactly that shape from any page it processes.

Does this tool send my JSON to a server?

No. Everything runs in your browser. Your JSON is never sent anywhere.

Does it handle null values?

Yes. A null value produces { "type": "null" }. If you need a field that accepts either a string or null, add anyOf manually after generating the base schema: [{ "type": "string" }, { "type": "null" }].

Why does the schema not include constraints like minimum or maximum?

Type inference can only tell you what type a value is, not what range it should be in. Whether a price can be negative or a rating must be between 1 and 5 is business logic that requires domain knowledge. The tool gives you the correct structural foundation. Adding value constraints is the next step and is done manually.

JSON Schema Generator: Infer Schema from JSON Instantly

JSON input

{
  "id": 1,
  "title": "Wireless Headphones Pro",
  "price": 149.99,
  "in_stock": true,
  "tags": ["audio", "wireless", "electronics"],
  "specs": {
    "battery_hours": 40,
    "weight_grams": 250,
    "color": "Midnight Black"
  },
  "reviews": [
    {
      "author": "Jane Doe",
      "rating": 5,
      "verified": true,
      "comment": "Best headphones I have owned."
    },
    {
      "author": "John Smith",
      "rating": 4,
      "verified": false,
      "comment": "Great sound, a bit heavy."
    }
  ]
}

$schema

·Mark all requiredDetect formats

JSON Schema

{
  // Generated JSON Schema will appear here...
}

How it works

Three steps from raw JSON to a typed schema.

01

Paste example JSON

Copy a real API response, a scraped page payload, or any JSON you already work with. The more representative the sample, the more accurate the generated schema.

02

Set your options

Choose your schema version, toggle format detection, and decide whether all fields should be marked required. Then click Generate Schema.

03

Copy or download

Copy the output to your clipboard or save it as a .json file. Paste it into the Spidra API schema parameter, an OpenAPI spec, or any JSON Schema validator.

What is JSON Schema?

JSON Schema is a standard vocabulary for describing the structure of JSON data. You define which fields a document should contain, what type each field holds, which fields are required, and what constraints the values must meet. Any tool or service that understands JSON Schema can then validate data against that description automatically.

It is used across the whole stack: OpenAPI specifications, ORM definitions, AI structured output from OpenAI and Anthropic, form validation libraries, and extraction APIs like Spidra. Write it once and it works as a contract in multiple places.

Why use it with AI scraping?

AI scrapers return loosely structured output unless you give them a contract. When you pass a JSON Schema to Spidra, it becomes the exact output specification. The AI knows which fields to extract, what type each one should be, and which are required. Every page comes back in the same shape, ready to store or pass downstream without any cleaning.

Generate your schema here, copy it, and paste it into the schema parameter in your Spidra API call. That is all it takes to go from a free-form webpage to typed, structured data at scale.

Why use this generator?

Accurate inference, smart merging, zero server calls.

Recursive type inference

Handles deeply nested objects, arrays of objects, mixed-type arrays, and null values. Integers and floating-point numbers are distinguished correctly at every level of nesting.

Smart array merging

When an array contains multiple objects, the tool merges all item schemas into one. Fields present in every object become required. Fields that appear only sometimes are kept as optional.

Draft 7 and 2020-12

Choose between Draft 7, which is compatible with OpenAI, Anthropic, and OpenAPI 3.0, and Draft 2020-12, the current standard used by OpenAPI 3.1 and modern validators.

Entirely client-side

Your data never leaves your browser. Schema generation runs locally in JavaScript with no server round trips and nothing stored anywhere.

Frequently asked questions

Everything you need to know about JSON Schema generation.

JSON Schema is a standard vocabulary for describing the structure of JSON data. You specify which fields a document should contain, what type each field holds, which fields are required, and what constraints apply. Validators, AI models, and API frameworks use it to check that data matches a known shape.

Start scraping for free.

Get 300 free credits to explore Spidra. Build your first scraper in minutes, not hours. Upgrade anytime as you scale.

We build features around real workflows. Usually within days.