Web Page Extractor

Hand over a URL and the fields you want; get back clean, structured JSON you can pipe into a spreadsheet, database, or another agent. No scraper to maintain.

What it does

• Reads structured signals first — JSON-LD, OpenGraph, meta tags — then visible content, surviving pages that break a selector.
• Matches your schema exactly — field names, types, nesting — or infers one from the page type (product, article, listing, company) and states the assumption.
• Normalizes values downstream tools expect: ISO dates, absolute URLs, prices split into value and currency.
• Returns null with a note for anything not on the page — never a guess.

What to send

One or more public URLs plus what to extract. Naming the fields or pasting a JSON schema gives the best results.

"From these product pages: name, price, and availability."

What you get back

One JSON object per URL — the fetched URL, a timestamp, your fields under data, and notes. Multiple URLs come back in one results array.

{
  "url": "https://example.com/p/widget-pro",
  "extracted_at": "2026-03-04T11:20:05Z",
  "data": {
    "name": "Widget Pro",
    "price": { "value": 129.00, "currency": "USD" },
    "availability": "InStock"
  },
  "notes": ["`rating` not on the page"]
}

Good for

• Turning product or listing pages into a price/spec table
• Pulling article metadata (title, author, date) at scale
• Harvesting contact details from about and team pages
• Feeding pipelines that need strict JSON

Limits

Public pages only — nothing behind a login or paywall it can't reach. Output comes verbatim from the fetched page; redirects report the final URL. It doesn't crawl beyond the pages you name. If it can't reach the web at all, you get one source_unavailable object, not a null-filled result that looks like your pages were empty.