Structured Data Validator + Rich Results & AEO Checker (Apify Actor)

Give it any public page URL and it detects every schema.org structured-data type on the page (both JSON-LD and microdata), validates each type against Google Rich Results requirements, and tells you exactly what's missing to win a rich result — plus an AEO (Answer-Engine Optimization) readiness read for AI search. HTML-only (no headless browser), fast, cheap, deterministic.

Built for SEO audits at scale, rich-result eligibility monitoring, schema QA in CI, competitor structured-data teardown, and AEO readiness checks.

What it validates

Per-type Google REQUIRED + RECOMMENDED property coverage for:

• Article / NewsArticle / BlogPosting, Product (merchant listing and product snippet), Offer / AggregateOffer, AggregateRating / Review, Recipe, Event, JobPosting, LocalBusiness (+ subtypes like Restaurant/Store), FAQPage, HowTo, BreadcrumbList, VideoObject, Organization, Course.

It encodes the real-world gotchas: HowTo is deprecated (detected for AEO but never a rich result), FAQPage rich results are deprecated (validated structurally only), Article & Organization have no Google-required props, Product splits into two experiences, conditional-required logic (offers ⇒ price), nested vs standalone ratings/reviews, BreadcrumbList needs ≥2 items, ratingValue must be in range, and microdata counts exactly like JSON-LD.

Input

{ "startUrls": [{ "url": "https://www.example.com/product/123" }], "maxConcurrency": 5, "maxPages": 100 }

maxPages capped at 200, maxConcurrency at 20.

Output — one STABLE record per URL (ok and error rows share the shape)

{
  "status": "ok",
  "requested_url": "...",
  "final_url": "...",
  "http_status": 200,
  "found": true,
  "complete": true,
  "page_type": "rich-eligible",
  "source": "json-ld",
  "detected_types": ["Product", "Offer", "AggregateRating", "BreadcrumbList"],
  "entities": [
    {
      "type": "Product",
      "source": "json-ld",
      "required_missing": [],
      "recommended_missing": ["offers.shippingDetails", "offers.hasMerchantReturnPolicy"],
      "rich_result_eligible": true,
      "errors": []
    }
  ],
  "rich_result_eligible_types": ["Product", "BreadcrumbList"],
  "validity_score": 88,
  "aeo_signals": {
    "has_faq": false,
    "has_howto": false,
    "has_breadcrumb": true,
    "has_article_meta": false,
    "answer_upfront": false
  },
  "fixes": [
    "Product.offers.priceCurrency missing — add ISO 4217 code for merchant listing eligibility."
  ],
  "extracted_at": "2026-06-04T..."
}

Field guide

• detected_types — every distinct schema.org @type found (JSON-LD + microdata, @graph and nested types flattened in, deduped, first-seen order).
• entities[] — one per recognized typed node: its required_missing, recommended_missing, rich_result_eligible, and hard errors (malformed date, rating out of range, non-numeric price, breadcrumb < 2 items, conditional-required miss).
• rich_result_eligible_types — types where at least one entity fully passes Google's REQUIRED props for a live rich result.
• validity_score — integer 0-100. REQUIRED-coverage dominates (80%), RECOMMENDED is bonus (20%), hard errors subtract. Capped at 79 until at least one type is rich-result eligible (so a page can't score "green" without a real, eligible type). 0 when nothing is found.
• aeo_signals — has_faq, has_howto, has_breadcrumb, has_article_meta, answer_upfront (answer-first content ≤300 chars) for Answer-Engine Optimization.
• fixes[] — concrete, prop-level remediation naming Type.property + why it matters.

found / complete semantics

• found = true iff ≥1 entity exists with a recognized non-empty schema.org @type (JSON-LD or microdata). A page with only <title>/OpenGraph, or only empty {} / @type-less nodes ⇒ found = false (never overbills).
• complete = true iff at least one type is rich_result_eligible. HowTo (deprecated) and bare FAQPage never make a page complete on their own.

Run locally / test

npm install
npm test     # 60 unit tests on the pure validator (node:test) — rich-result rules + edge cases

Publish to Apify (account-holder's step)

$npm install -g apify-cli && apify login && apify push

Notes / safety

• SSRF-guarded, robots-respecting, rate-limited, cost-capped (shared src/lib/actor_runner.js).
• Stores only derived validation results — no raw page bodies.
• HTML-only: client-rendered pages that inject JSON via JS return found:false with render_required:true. Core logic in src/validate.js (pure, deterministic, unit-tested).
• Validation rules web-verified against developers.google.com/search/docs structured-data docs (June 2026).