Schema Markup Scraper API - JSON-LD, Microdata & RDFa

Extract structured-data and SEO metadata from public web pages: JSON-LD, HTML Microdata, basic RDFa, schema.org types, canonical URLs, Open Graph, Twitter Cards, and meta tags. Every row also includes the final URL, HTTP status, content type, retrieval time, analyzed-byte count, truncation state, and an interpretation warning.

Pricing: $0.002 per processed URL plus the memory-scaled Actor start event ($0.0001 at the default reservation). The one-page default costs at most $0.0021. Buyer charge limits are applied before URLs are fetched.

Why use it

• Audit structured-data coverage across landing pages, product pages, articles, and local-business pages.
• Export parsed JSON-LD instead of writing site-specific HTML parsers.
• Compare schema.org types and social metadata across competitors.
• Find malformed JSON-LD blocks and pages where static HTML exposes no structured data.
• Feed source-linked, point-in-time metadata into SEO QA or content pipelines.

Input

Field	Type	Default	Description
urls	string array	https://www.apple.com	Public HTTP/HTTPS pages. One paid output item per processed URL.
includeJsonLd	boolean	true	Parse application/ld+json scripts and @graph arrays.
includeOpenGraph	boolean	true	Extract og:* properties.
includeTwitterCard	boolean	true	Extract twitter:* properties.
includeMetaTags	boolean	true	Extract HTML meta name/content pairs.
concurrency	integer	5	Parallel requests, from 1 to 20.
timeout	integer	30	Per-URL timeout in seconds, from 5 to 120.
maxResults	integer	1	Total URL/result cap, up to 10,000. A buyer charge limit can reduce it further.

Microdata and RDFa evidence are always extracted because they are distinct schema-markup formats, not JSON-LD options.

Output

Each requested URL produces one stable 23-key row. Arrays and objects are empty rather than null when evidence is absent.

{
  "url": "https://www.apple.com",
  "finalUrl": "https://www.apple.com",
  "httpStatusCode": 200,
  "contentType": "text/html; charset=utf-8",
  "fetchedAt": "2026-07-25T05:13:44Z",
  "analyzedBodyBytes": 249240,
  "bodyTruncated": false,
  "title": "Apple",
  "canonicalUrl": "https://www.apple.com/",
  "jsonLd": [
    {
      "@context": "https://schema.org",
      "@type": "Organization",
      "name": "Apple"
    }
  ],
  "microdata": [],
  "rdfa": [],
  "openGraph": {
    "title": "Apple",
    "type": "website",
    "url": "https://www.apple.com/"
  },
  "twitterCard": {},
  "metaTags": [
    {"name": "description", "content": "..."}
  ],
  "schemaTypes": ["ContactPoint", "Organization", "SearchAction", "WebPage", "WebSite"],
  "microdataTypes": [],
  "rdfaTypes": [],
  "allSchemaTypes": ["ContactPoint", "Organization", "SearchAction", "WebPage", "WebSite"],
  "malformedJsonLdBlocks": 0,
  "hasStructuredData": true,
  "interpretation": "Point-in-time static HTML evidence only. Client-rendered or blocked markup can be missed; extracted syntax is not Google rich-result eligibility or semantic-validity certification.",
  "error": ""
}

Format details

• schemaTypes recursively includes nested JSON-LD types such as Product and Offer.
• microdata contains each itemscope, its declared type/ID, and directly owned itemprop values.
• rdfa provides basic typeof, about, and property extraction; it is not a complete RDF graph expansion.
• allSchemaTypes is the sorted union of JSON-LD, Microdata, and RDFa types.
• malformedJsonLdBlocks counts script blocks that are present but not valid JSON.
• Non-2xx pages remain explicit paid evidence rows with httpStatusCode and a non-empty error instead of silently looking successful.

Evidence limits

• This Actor inspects the first 5 MiB of static response HTML. It does not execute JavaScript. Client-rendered markup may therefore be absent.
• WAFs, consent pages, geolocation, authentication, personalization, and bot filtering can change the observed response.
• Parsed syntax does not certify schema.org semantic validity, Google rich-result eligibility, search ranking, canonical selection, ownership, or current indexing.
• Basic Microdata/RDFa extraction is designed for practical auditing, not standards-complete RDF processing.
• Private, loopback, link-local, non-HTTP, and malformed literal URLs are rejected. Only use publicly accessible pages you are authorized to request.
• Redirects are followed up to 10 hops; url preserves the request and finalUrl records the observed destination.