Welcome to Refyne

Refyne is an LLM-powered API that transforms unstructured websites into clean, structured data. Just describe what you want in plain English, or define a precise schema - either way, you get typed JSON output.

Quick Start

Extract data from any webpage using natural language:

curl -X POST https://api.refyne.uk/api/v1/extract \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://demo.refyne.uk/products/5",
    "schema": "Extract the product name, brand, price, and whether it is in stock"
  }'

{
  "data": {
    "product_name": "Red Nail Polish",
    "brand": "Nail Couture",
    "price": 8.99,
    "in_stock": true
  }
}

Need more control? Use a typed schema instead:

curl -X POST https://api.refyne.uk/api/v1/extract \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://demo.refyne.uk/products/5",
    "schema": {
      "name": "string",
      "price": "number",
      "in_stock": "boolean"
    }
  }'

Features

• Natural Language or Schemas - Describe what you want in plain English, or use typed schemas for precise control
• Single Page Extraction - Extract structured data from any webpage
• Multi-Page Crawling - Follow links and extract data at scale
• Multiple Output Formats - JSON, JSONL, or YAML to fit your workflow
• Multiple LLM Providers - OpenRouter, Anthropic, OpenAI, or bring your own
• Site Analysis - Automatically detect page structure and suggest schemas

Next Steps