How fb reads Facebook

fb does not log in. It reads the same server-rendered pages Facebook serves to search engines: the public version of a Page, profile, group, or post, with the text, counts, media, and a few preview comments baked into the HTML. There is no cookie, no password, and no browser to drive.

fb whoami

MODE       USER_AGENT
anonymous  Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)

What you get

For a public Page, profile, group, or post, fb returns the same fields you would read on the page itself:

• Page and profile records: name, about text, like and follower counts, avatar.
• Posts: text, creation time, reaction, comment, and share counts, attached media, and outbound links.
• A handful of preview comments per post, with the commenter's name.

fb id classifies any URL or id offline and never touches the network.

The trade-off: depth

Reading the crawler surface keeps fb to a single binary with no login, at the cost of depth. Two limits are worth knowing:

• Recent posts only. A feed exposes roughly the most recent posts, not the full history. There is no deep pagination, so --limit above what the page carries simply returns what is there.
• Preview comments only. Each post carries a few preview comments, not the full thread, and the commenter attribution is approximate. Comment timestamps are not exposed on this surface.

What stays private

Anything Facebook does not put on the public crawler page is not reachable: the content behind a login wall, private groups and profiles, full comment threads, and per-reactor lists. When a target is private or removed, fb exits 4 (login wall) or 3 (content unavailable) rather than guessing.

Staying polite

fb defaults to a two-second delay between requests (--rate) and caches responses on disk, so re-runs do not hit Facebook again. Raising the rate or running many workers risks a temporary block, which fb surfaces as exit 5. Fetch at a human pace.