New: monitor which AI crawlers actually visit your site with Sona Agent Analytics  |  See the platform →

FacebookBot

Meta's legacy training crawler, no longer listed on Meta's current crawler page. Superseded by meta-externalagent.

Reference reviewed by Sona

This token is deprecated or superseded. Read the detail below before relying on a rule that names it - it may no longer match live traffic.

OperatorMeta
PowersMeta speech and language model training
PurposeModel training
User-agent tokenFacebookBot
Respects robots.txtYes

FacebookBot crawled public web content to help train Meta's language technology, with speech recognition specifically called out in its original documentation. It predates Meta's current crawler family and was the token most robots.txt files named when they wanted to address Meta at all.

It no longer appears on Meta's web crawlers documentation page, which now lists facebookexternalhit, Meta-WebIndexer, Meta-ExternalAds, Meta-ExternalAgent, and Meta-ExternalFetcher. Treat FacebookBot as legacy: meta-externalagent is the current training token, and a rule naming only FacebookBot almost certainly does not do what its author intended any more.

Keeping the old rule is harmless, and there is a mild argument for it while any residual infrastructure still uses the identifier. But it is not a substitute for rules covering Meta's four current tokens - and a training opt-out that names FacebookBot alone is effectively no opt-out.

One long-standing confusion is worth settling: FacebookBot is not facebookexternalhit. The latter generates link previews when someone shares your URL, and Meta notes it may bypass robots.txt for security and integrity checks. Blocking FacebookBot has never affected how your links unfurl on Facebook.

How FacebookBot behaves

  • Absent from Meta's current crawler documentation page; superseded by meta-externalagent.
  • Originally documented as supporting language and speech model training.
  • Distinct from facebookexternalhit - blocking it never affected link previews.
  • Unverifiable and undocumented, which makes present-day traffic under this name questionable.

Full user-agent string

FacebookBot/1.0 (+https://developers.facebook.com/docs/sharing/webmasters/facebookbot/)

Allow FacebookBot

Costs nothing in practice, since the token is largely historical - and any residual crawl still feeds Meta's language and speech work.

User-agent: FacebookBot
Allow: /

Block FacebookBot

Tidies up a legacy training token - but pair it with meta-externalagent, which is where Meta's training crawl actually happens now.

User-agent: FacebookBot
Disallow: /

How to verify FacebookBot

No published verification method

Meta has never published IP ranges or a reverse-DNS convention for its crawlers, and this token is no longer documented at all, so verification is impossible. That combination - a recognizable name, no verification, and no current documentation - makes FacebookBot an attractive identity for unrelated scrapers. Traffic under this user-agent today deserves more suspicion than traffic under Meta's current tokens.

Check an IP against this bot

Commonly confused with FacebookBot

Meta-ExternalAgent

The current training token. If your goal is opting out of Meta's training crawl, this is the rule that matters - FacebookBot alone will not do it.

Meta-WebIndexer

Handles Meta AI search indexing and citations, a role that did not exist when FacebookBot was Meta's only documented crawler.

FacebookBot FAQs

Is FacebookBot still active?

It is no longer listed on Meta's crawler documentation page, which now covers facebookexternalhit, Meta-WebIndexer, Meta-ExternalAds, Meta-ExternalAgent, and Meta-ExternalFetcher. Treat it as a legacy token and write rules for the current ones.

Will blocking FacebookBot break link previews on Facebook?

No. Previews come from facebookexternalhit, a separate token that Meta notes may bypass robots.txt for security and integrity checks. The two have always been independent.

I block FacebookBot. Am I opted out of Meta AI training?

No. Meta's training crawl now runs under meta-externalagent. A rule naming only FacebookBot is effectively no opt-out at all.

Should I remove the FacebookBot rule?

No need - it is harmless and may still catch residual traffic. Just do not treat it as coverage. Add rules for Meta's four current tokens alongside it.

Can FacebookBot read your page right now?

Test any URL and see exactly what AI crawlers receive.

Check my site