Facebook Public Page Stats Scraper

Scrapes public Facebook page metadata — page name, like count, "talking about this" count, description and profile image. No login, no API key, no access token.

Useful for brand-size benchmarking, competitor tracking, and enriching a company database with social reach figures.

What makes this different

talking_about_count is the metric that actually moves. Like counts are a lifetime total that barely changes week to week; "talking about this" is Facebook's own recent-engagement figure, and it's the one that separates a page with real current activity from a page with a big historical following and nothing happening. Both are returned as clean integers, parsed out of the description text where Facebook buries them.

A clean description, plus the original. Facebook prepends the page name and the stats sentence to og:description ("NASA - ... . 28,711,768 likes · 124,772 talking about this. Explore..."). Passing that straight through duplicates fields you already have. description is the page's own blurb with that prefix stripped; raw_description keeps the untouched original so nothing is lost.

A page that isn't publicly viewable fails clearly. Facebook returns HTTP 200 for a nonexistent page — the only tell is that the title becomes "Log in to Facebook" and the counts vanish. That's detected explicitly, so you get a stated failure instead of a row full of nulls that looks like a real page with no engagement.

What's honestly out of scope

No posts. This Actor returns page-level stats only. A Facebook page served to a logged-out client contains no post feed at any depth — verified directly: zero occurrences of the feed's own data markers in the HTML, on both the desktop and mobile hosts. The legacy no-JS host now redirects to a login page outright. Anything claiming to return logged-out page posts is doing something this Actor does not do.

Input

Field	Type	Notes
pages	array	Page names (the part after facebook.com/) or full page URLs.
proxyConfiguration	object	Defaults to Residential.

Output

{
  "page_id": "facebook.com:nasa",
  "name": "NASA - National Aeronautics and Space Administration",
  "like_count": 28711768,
  "talking_about_count": 124772,
  "description": "Explore the universe and discover our home planet. There's space for everybody.",
  "url": "https://www.facebook.com/nasa/"
}

were_here_count is present only for pages with a physical location. Each run also writes a RUN_COVERAGE record to the key-value store with what was requested, what came back, and any per-page failures.

Local development

pip install -r requirements.txt
python test_local.py nasa cocacola netflix --out sample_output.json
python test_local.py https://www.facebook.com/nasa

sample_output.json is real output from a live run.