# Structured Data Extraction (JSON Schema)

Structured data extraction from a web page or your own text: send a JSON Schema, get JSON that is validated against it.

Send a URL and your own JSON Schema: prices, specs, contact details, article facts, whatever the page holds. Every answer is validated against the schema before it reaches you, and a page that cannot be made to fit costs nothing. Already have the HTML or text? Send it as content instead of a URL. No browser is run, so pages that need JavaScript or a login are refused. We honour robots.txt and opt-outs and ask each site slowly. Do not send personal data.

- You send your own JSON Schema, and every answer is validated against it before it reaches you.
- A page that cannot be made to fit costs nothing.
- Already have the HTML or text? Send it as content instead of a URL.

## Calls

### `POST https://readymadeapi.com/v1/web-extract`

Read one public page, or content you send, and fill your JSON Schema from it. The output is checked against the schema; if it does not validate, nothing is charged.

- Price: $0.012 per page
- MCP tool: `web_extract_extract`
- Answers in the same call.

Request:

```json
{
  "content": "Oak desk Linnea, 160 x 80 cm. Solid oak top, steel legs. Now 449,00 EUR, was 529,00 EUR. In stock, ships in 3 days. SKU LN-160-OAK.",
  "schema": {
    "properties": {
      "currency": {
        "type": "string"
      },
      "in_stock": {
        "type": "boolean"
      },
      "name": {
        "type": "string"
      },
      "price": {
        "type": "number"
      },
      "sku": {
        "type": "string"
      }
    },
    "required": ["name", "price", "currency", "in_stock"],
    "type": "object"
  }
}
```

Answer:

```json
{
  "data": {
    "currency": "EUR",
    "in_stock": true,
    "name": "Oak desk Linnea, 160 x 80 cm",
    "price": 449.0,
    "sku": "LN-160-OAK"
  },
  "url": null,
  "truncated": false,
  "validated": true
}
```

## Limits

- One page or one piece of content per call.
- The first 30,000 characters of text are read and the answer says when the page was cut. Content up to 1 MB, schema up to 16 KB.

## Questions

**Does it get past logins, CAPTCHAs or bot protection?**

No. No browser is run, and we honour robots.txt and opt-outs. Pages that need JavaScript or a login are refused and not charged.

**May I send personal data?**

No. The text goes to a language model at an outside provider, on a route with zero data retention switched on.

**What is it for?**

Prices, specifications, contact details and facts from pages you are allowed to read, in exactly the shape your code expects.

**Why did I get a 429 site_busy?**

We ask each site slowly, across all our users. When a site has been asked too often you get an uncharged 429 with a Retry-After.

## Using it

- Send `Authorization: Bearer <key>` and a JSON body. Keys and credit: https://readymadeapi.com/account
- Every answer carries `X-Request-Id`, `X-Units` and `X-Charged-Usd`.
- An error is a JSON problem with a `title` to act on and a `detail` in plain words; nothing is charged for it.
- Full schema: https://readymadeapi.com/openapi.json
