Skip to content
CiteScore

Guide

How to get cited by Meta AI

What Meta publishes for Meta AI: Meta-WebIndexer is the crawler Meta ties to citations. Meta-ExternalAgent is a separate training and indexing crawl. Meta does not publish a citation rate.

18 October 2026 · CiteScore Team · 11 min read

Meta publishes a crawler page for the user agents that fetch other people’s sites. That page is the public rule for how Meta prefers to be told yes or no. It is not a ranking formula. Meta does not publish a citation rate, a required schema type, or a word count. CiteScore does not log into Meta AI, WhatsApp, Instagram, Facebook, or meta.ai. The claims below are from Meta’s web crawlers documentation, read on 18 October 2026.

Meta AI answers in WhatsApp, Instagram, Facebook, and on meta.ai. Meta’s crawler page describes one set of crawlers for Meta AI responses. It does not publish a separate robots.txt token for each of those products, and it does not say that a page allowed for one crawler is quoted in every app. The same kind of split — a published control, and the files the operator does not require — is written up for ChatGPT, Perplexity, Claude, Gemini, Microsoft Copilot, Grok, DeepSeek, Mistral, Brave Search, and You.com, and DuckDuckGo, and Perplexity Comet. The short map is how to get cited by AI answer engines.

Five crawlers, and only one is tied to a citation

The page names five user agents. The product token in a heading is not always the spelling in robots.txt. Meta’s own example uses the lowercase token meta-externalagent. robots.txt matching is case-insensitive. The groups below use the spelling Meta publishes in that example and in the user-agent strings.

TokenWhat Meta says it does
meta-webindexerNavigates the web to improve Meta AI search result quality. Meta says allowing Meta-WebIndexer in robots.txt helps Meta cite and link to your content in Meta AI’s responses. The user-agent string begins meta-webindexer/1.1.
meta-externalagentCrawls for uses such as training foundation models or improving products by indexing content directly. Meta does not describe this token as the citation switch. The string begins meta-externalagent/1.1.
meta-externalfetcherFetches individual links a person requested, including agentic tasks. Meta says this crawler may bypass robots.txt. A disallow is not a lock.
facebookexternalhitBuilds the title, description, and thumbnail when a link is shared on Facebook, Instagram, or Messenger. Meta says it might bypass robots.txt for security or integrity checks, such as looking for malware. That is a preview fetch, not a Meta AI citation.
meta-externaladsCrawls for advertising and other business products. Meta does not tie this token to citations in Meta AI answers.

Allow the crawler Meta names for citations

A crawler that honors robots.txt uses the group that names its token. It uses User-agent: * only when it has no group of its own. If the wildcard group already allows the host, you do not need a second group to allow Meta-WebIndexer. If the wildcard group is Disallow: /, add a group that names the token or the disallow applies to it.

User-agent: meta-webindexer
Allow: /

User-agent: meta-externalagent
Allow: /
Disallow: /private/

The first group is the citation request: let Meta-WebIndexer fetch the host. The second group is Meta’s own example for Meta-ExternalAgent, which allows the host and disallows one private path. Copy any private path you still want blocked into a named group. A named group does not inherit the wildcard rules. Meta says to allow up to 24 hours for a robots.txt change, because crawlers may cache the file for that long.

Allowing Meta-WebIndexer does not require a decision about training. Meta describes Meta-ExternalAgent separately, and a group for one token does not write the other. Meta does not say the two share a crawl, and it does not say that blocking Meta-ExternalAgent removes a page from Meta AI answers. Treat them as two requests. Meta does not say Meta-WebIndexer or Meta-ExternalAgent bypass robots.txt. It does say Meta-ExternalFetcher may, because a person asked for the fetch, and that FacebookExternalHit might, for a security check. robots.txt is not a firewall. A WAF can still return 403 after the file says allow.

The wider token list, including Meta-ExternalAgent as it was recorded for the reference table, is the AI crawler user-agent reference. The five tokens on the robots.txt guide are GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and CCBot. Meta’s tokens are not in that list.

FacebookBot is not on the current page

Older publisher notes sometimes tell you to allow or block FacebookBot for Meta’s AI crawl. Meta’s crawler page, read on 18 October 2026, does not list that token. It lists the five above. This guide does not treat FacebookBot as the control for Meta AI citations, and it does not invent a robots.txt behavior for a name the page does not document. If your logs show a client calling itself FacebookBot, that string can be spoofed. Confirm it the way you would confirm any other user agent before you write a rule for it. If Meta later publishes FacebookBot and says what a disallow does, that documentation replaces this paragraph.

Put the entity in the HTML

Meta’s page describes a fetch of public pages. It does not describe a preferred heading length. Put the site name, what the site does, and who it is for in the HTML of the homepage, in the title, the first paragraph, and the footer, using the same name in each place. A fact that exists only in an image, or only after JavaScript runs, is a fact many crawlers never read. Meta does not say its crawlers execute JavaScript. Text in the server HTML still travels either way.

WebSite JSON-LD repeats that name, the canonical URL, an alternate name the page already uses, and the publisher. Organization JSON-LD repeats the publisher’s logo and profiles. Meta does not require either script. They are labels for a parser, not a switch on meta.ai. The WebSite schema generator and the Organization schema generator copy those fields in the browser. The field list, and what to leave out, is the schema guide.

Answer the question in the first paragraph

A page that is fetched still has to contain the sentence an answer could quote. State the answer before the navigation, the pricing table, or a story about the company. Use the question a person would ask as a heading when that question is the point of the page. Meta does not publish a word count for that sentence, and this page does not invent one. Specific names, dates, and the source of a number are easier to attribute than a slogan.

Third-party mentions are not a documented switch

Meta’s crawler page does not say that a Wikipedia article, a review, or a directory listing causes a citation in Meta AI. A page someone else publishes can be fetched on its own. That is a separate URL, not a ranking factor Meta has written down. What you can control is consistency: the same name on your homepage and on profiles you actually own, linked with sameAs only when the profile is yours. Two sites that describe the same product under two names give a model two entities to choose from. Aligning the name is entity clarity. It is not a citation guarantee.

Free AI Answer Snapshot, then the Fix Pack

The free AI Answer Snapshot asks logged-out ChatGPT and Perplexity who they name for a local business and sends one page. AI answers vary run to run. No payment. The Fix Pack, $49 is the done-for-you work after you have seen who was named.

The free check reports GPTBot, PerplexityBot, and Google-Extended. It does not read meta-webindexer, meta-externalagent, or FacebookBot. A clear row on that check is not a Meta AI citation. The GEO audit scores one public URL from the HTML we fetch and returns about ten fixes. It does not open WhatsApp, Instagram, Facebook, or meta.ai.

Run the free check on the host. It reads the homepage, robots.txt for GPTBot, PerplexityBot, and Google-Extended, /llms.txt, a sitemap, and JSON-LD on the homepage. It does not log into an answer engine. Paid files are listed at checkout.

Questions about Meta AI citations

Does allowing Meta-WebIndexer guarantee a citation in Meta AI?

No. Meta’s crawler page says allowing Meta-WebIndexer helps Meta cite and link to your content in Meta AI’s responses. It does not publish a citation rate, a required schema type, or a word count, and it does not say that an allowed page is quoted in WhatsApp, Instagram, Facebook, and on meta.ai. An allow is permission to fetch. The mention is still the model’s decision after a fetch.

Is FacebookBot the robots.txt token for Meta AI?

Not on the crawler page read for this guide. On 18 October 2026 that page lists FacebookExternalHit, Meta-WebIndexer, Meta-ExternalAds, Meta-ExternalAgent, and Meta-ExternalFetcher. It does not list FacebookBot. A group written for a name Meta does not publish is not the control on that page. If Meta later documents FacebookBot and says what disallowing it does, that page becomes the source.

Does blocking Meta-ExternalAgent remove a site from Meta AI answers?

Meta does not say that. Meta-ExternalAgent is the crawler Meta describes for training foundation models or improving products by indexing content directly. Meta-WebIndexer is the crawler Meta ties to citations in Meta AI responses. A robots.txt group for one token does not write the other. Meta does not say Meta-WebIndexer ignores robots.txt. Meta does say Meta-ExternalFetcher may bypass the file because a person requested the fetch, and that FacebookExternalHit might bypass it for security checks.

Does Meta require WebSite schema or an llms.txt file?

No. Meta’s crawler page does not list a schema type or an llms.txt file. WebSite and Organization JSON-LD repeat the name and URL a person can already see. They are not a Meta switch. The free check reads GPTBot, PerplexityBot, and Google-Extended. It does not read meta-webindexer, and it does not log into Meta AI.

Sources

Request a free AI Answer Snapshot

Same request as the free snapshot page. AI answers vary run to run. This is a snapshot, not a ranking, and not a promise that any assistant will name the business.

All fields are required.

We email the snapshot to this address. We do not ask for passwords, cookies, or content behind a login. We do not sell emails. Privacy.

Email this request instead
citescore@mail.grokbot.com

No mail app opened? Copy this and email it to citescore@mail.grokbot.com from any inbox.

If that send fails, email this request instead, or use the copy buttons.

Sends the business name, city, service, website, and email to citescore@mail.grokbot.com through FlowForm. This page path is included. No account and no paid checkout.

Prepared by the CiteScore Team.