Skip to main content

​
Overview

The json_schema option lets you define a custom extraction schema so the parser returns only the structured fields you specify. This is powerful for extracting specific data points from any web page without writing custom parsing logic.

​
Schema Format

The schema is an object with a properties array. Each property defines a field to extract: 
{
  "properties": [
    {
      "name": "field_name",
      "type": "string"
    }
  ]
}

​
Supported Types

TypeDescription
stringText content
numberNumeric values
booleanTrue/false values
mapNested object — use properties for nested fields
arrayList of items — use items to define the item schema

​
Examples

​
Extract Article Metadata

{
  "url": "https://example.com/blog/article",
  "options": {
    "json_schema": {
      "properties": [
        { "name": "title", "type": "string" },
        { "name": "author", "type": "string" },
        { "name": "publish_date", "type": "string" },
        { "name": "word_count", "type": "number" }
      ]
    }
  }
}

​
Extract Product Data

{
  "url": "https://example.com/product/12345",
  "options": {
    "json_schema": {
      "properties": [
        { "name": "product_name", "type": "string" },
        { "name": "price", "type": "number" },
        { "name": "in_stock", "type": "boolean" },
        {
          "name": "reviews",
          "type": "array",
          "items": {
            "properties": [
              { "name": "rating", "type": "number" },
              { "name": "text", "type": "string" },
              { "name": "author", "type": "string" }
            ]
          }
        }
      ]
    }
  }
}

​
Nested Objects

{
  "url": "https://example.com/company",
  "options": {
    "json_schema": {
      "properties": [
        { "name": "company_name", "type": "string" },
        {
          "name": "contact",
          "type": "map",
          "properties": [
            { "name": "email", "type": "string" },
            { "name": "phone", "type": "string" },
            { "name": "address", "type": "string" }
          ]
        }
      ]
    }
  }
}

​
Notes

  • JSON schema extraction works with any URL, not just supported platforms
  • The parser uses AI to intelligently match your schema fields to page content
  • Field names should be descriptive — the parser uses them to understand what data to extract
  • JSON schema extraction costs 1 token (standard) unless combined with AI options