Facebook Business Contact Extractor

Pull the public contact details businesses put on their own Facebook pages: email, phone, website, street address, category, price range, services, opening status, follower count and rating.

Feed it a list of pages, or let it find the pages by keyword and country.

No login, no cookies to paste, no proxy required.

How keyword discovery works, and why

Facebook's own page search is closed to signed-out clients. /search/pages/, /search/places/, /search/top/ — every vertical returns HTTP 404. There is no public "find pages matching X" endpoint any more.

The Ad Library is the exception: it is keyword-addressable, it is country-scoped, and every result names its advertiser's page. So that is where discovery runs. One useful side effect: every business it finds is currently spending money on advertising, which is a sharper filter than a keyword match.

A live run for "dentist" in the US surfaced 53 distinct business pages in one pass.

If you already have a list, skip discovery and pass pages instead.

What you get

Field	Example
pageName, primaryCategory	BARBUTO | New York NY, Italian restaurant
email, phone	info@barbutonyc.com, +1 212-924-9700
website	http://www.barbutonyc.com/ (unwrapped from Facebook's redirect)
address	113 horatio street, New York, NY, United States, 10014
priceRange, services	Price range · ££, ["Dine in"]
hoursLabel	Closed now
followers, recommendPercent, reviewCount	6987, 94.0, 1236
contactScore	4 — how many of the four channels are published
discoveredVia, discoveryQuery	ad-library, dentist @ US

With includeTransparency on, each page also gives pageCreationDate (17 February 2010) and isRunningAds — useful for telling an established business from a week-old one, and a prospect already buying ads from one that is not.

Input example

{
  "searchKeywords": ["dental implants", "orthodontist"],
  "countries": ["US", "GB"],
  "maxPages": 300,
  "requireAnyContact": true,
  "includeTransparency": true,
  "exportFormats": ["xlsx"]
}

Or, with a list you already have:

{
  "pages": ["BarbutoNYC", "https://www.facebook.com/gramercytavern"],
  "requireEmail": true
}

Honest limits

• Opening hours are not published. Facebook shows logged-out visitors only Open now / Closed now — there is no weekly schedule in the payload. The field is named hoursLabel rather than hours so nobody mistakes it for one.
• A null contact field means the business publishes none, not that the parser missed it. That distinction is the whole point of a lead list, so the actor never falls back to guessing a value from nearby text. Addresses found in the page's description rather than its contact fields are reported separately as emailsFromDescription.
• Missing slugs are reported as errors. Facebook answers HTTP 200 for pages that do not exist, so a status-code check would return a blank row for every typo. This actor checks the payload size and gives you a page_not_found error row instead.
• These pages are big — 8 to 18 MB each. That is why concurrency defaults to 4. Raising it speeds a run up and costs memory and bandwidth.
• Follower counts are abbreviated by Facebook; the parsed number is approximate and the raw label is kept beside it.