Local Business Directory Extractor (single page)

Extract structured business data from a single public business-listing or company-profile page — deterministically, with no proxy, no headless browser, no AI and no paid API.

What it does

Give it one business/company page URL. It fetches that page once and parses the business data out of the page's schema.org LocalBusiness / Organization JSON-LD, OpenGraph tags and meta/visible-text heuristics, returning a single structured record:

url, status, name, address, phone, email, website, hours[], rating,
review_count, categories[], geo{lat,lng}, description, method,
parse_confidence, extracted_at, error

This is single-page extraction, not bulk directory scraping (see Limitations).

Input

field	type	notes
url	string	a single public business/company page URL
urls	array of string	optional bounded batch, each handled like url (cap 50)

Example input:

{
  "url": "https://www.apple.com/"
}

Output

One dataset record per URL. Example (fields populated depend on what the page publishes):

{
  "url": "https://www.example-restaurant.com/",
  "status": "completed",
  "name": "Joe's Diner",
  "address": "123 Main St, Springfield, IL, 62704, US",
  "phone": "+1-217-555-0123",
  "email": "hello@joesdiner.com",
  "website": "https://www.example-restaurant.com/",
  "hours": ["Mo-Fr 08:00-22:00", "Sa Su 09:00-23:00"],
  "rating": "4.6",
  "review_count": "812",
  "categories": ["Restaurant", "FoodEstablishment"],
  "geo": { "lat": "39.7817", "lng": "-89.6501" },
  "description": "Classic American diner serving breakfast all day.",
  "method": "jsonld_business",
  "parse_confidence": "high",
  "extracted_at": "2026-06-24T10:00:00+00:00",
  "error": null
}

When extraction fails, the record is still returned with status (failed / blocked / empty) and a human-readable error; the actor never raises.

Use cases

• Enrich a CRM / lead record from a company's public profile page (name, phone, address, hours, geo).
• Pull canonical NAP (Name / Address / Phone) data for a single known business listing.
• Verify or normalize the structured LocalBusiness data a site publishes.
• Spot-check whether a business page exposes valid schema.org markup.

How it works (deterministic, code-only)

Extraction is pure code, tried most-reliable-first:

1. JSON-LD / schema.org LocalBusiness (any subtype) / Organization — name, address (PostalAddress), telephone, email, url, openingHours / openingHoursSpecification, aggregateRating, geo, categories / subtypes. Highest-confidence path.
2. OpenGraph / business meta tags — og:site_name, og:title, og:description, og:url, business:contact_data:*, place:location:*.
3. Plain <meta> / <title> / <h1> + visible-text heuristics — mailto: / tel: links and phone / email regex fallback.

The layer used is reported in method, and a code-owned parse_confidence (high / medium / low / none) is attached to every record.

Cost-safety ($0 idle, $0 uncovered per run)

• No proxy — direct bounded HTTP GET.
• No headless browser — static HTML fetch only.
• No AI / LLM — pure deterministic parsing.
• No paid third-party API.

The only cost is Apify platform compute for the run itself.

Always-on security (SSRF-guarded, fail-closed)

• Private / loopback / link-local / reserved IPs are blocked (SSRF guard), re-validated on every redirect hop.
• A domain blocklist rejects login-walled / ToS-restricted directory sites (LinkedIn, Facebook, Yelp, Yellow Pages, TripAdvisor, Glassdoor, ...).
• Hard caps: 5s connect / 10s read timeout, 2 MB body, 3 redirects.

Limitations (honest)

• Single page, not bulk crawling. It extracts one business per URL. It does not spider a directory, follow result links, or paginate through listing pages — that would require proxies and carries ToS / legal risk.
• It reports what the page publishes. If a page has no LocalBusiness / Organization JSON-LD and no useful OpenGraph/contact markup, fields will be sparse and parse_confidence will be low / none. The actor does not invent values.
• Blocklisted directories are refused by design (returned as blocked), to respect their terms of service.