AI ANSWER ENGINE
What is Meta AI?
Meta AI is an AI answer engine run by Meta Platforms, Inc. Where its answers come from is not stated. On record: Bing, per a September 2023 statement; no later Meta page read confirms it. On structured data, this research records the stance “Its pages say nothing about it”.
Last checked November 10, 2026 · 5 sources
- Type
- AI answer engine
- Run by
- Meta Platforms, Inc.
- Results come from
- Not statedAlso on record: Bing, per a September 2023 statement; no later Meta page read confirms it
- Recorded markets
- Australia, Canada, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, United States of America, Zambia, Zimbabwe
- Where its AI chatbot share is highest
- StatCounter does not track Meta AI in its AI chatbot figures.
- AI chatbot share, worldwide
- StatCounter does not track Meta AI in its AI chatbot figures.
- Structured data formats
- None documented
- Rich result types
- None on record
Where does Meta AI get its results?
Not stated for the current product. Meta-WebIndexer crawls the web to improve Meta AI search results[1], which points to a Meta-built index, but no Meta page read says where answers come from. In 2023 Meta said Meta AI gets real-time information in text chats "through our search partnership with Bing"[2]. The 2024 post[3] and the help pages[4, 5] name no provider, so whether the Bing arrangement is current is not confirmed.
Engines and providers it names
BingOld statement, not confirmed since
Bing, per a September 2023 statement; no later Meta page read confirms it[2]
Conditions recorded: Statement dated 2023-09-27, for real-time information in text chats. No later Meta page read names a provider, so current status is not confirmed.
Words on the source page: “through our search partnership with Bing”
Engines that use results from Meta AI
The research records no engine that takes its general results from Meta AI or is built on it by the same company.
Which countries use Meta AI?
StatCounter does not track Meta AI in its AI chatbot figures. None of the labels in its AI chatbot figures for October 2026 maps to Meta AI in this research. StatCounter Global Stats
Recorded markets
The research records Australia, Canada, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, United States of America, Zambia and Zimbabwe as the main markets for Meta AI.
Does Meta AI use structured data?
Stance recordedIts pages say nothing about it
The Meta crawlers page mentions Open Graph only for link previews and does not mention schema.org or any other structured data. The Meta AI help pages[4, 5] are silent on markup.[1]
| Format | Documented support | What the research records |
|---|---|---|
| JSON-LD | Not documented | Not mentioned for Meta AI on the Meta crawlers page[1] or the Meta AI help pages[4, 5]. Not documented is not the same as not supported.[1] |
| Microdata | Not documented | Not mentioned for Meta AI on the Meta crawlers page[1] or the Meta AI help pages[4, 5]. Not documented is not the same as not supported.[1] |
| RDFa | Not documented | Not mentioned for Meta AI on the Meta crawlers page[1] or the Meta AI help pages[4, 5]. Not documented is not the same as not supported.[1] |
| Open Graph | Not documented | The crawlers page ties Open Graph only to the link preview crawler FacebookExternalHit (properties must sit within the first 1 MB of the page). That is link sharing. The page does not say Meta AI reads Open Graph, and the Meta AI help pages[4, 5] do not mention it. Not documented is not the same as not supported.[1] |
| Microformats | Not documented | Not mentioned for Meta AI on the Meta crawlers page[1] or the Meta AI help pages[4, 5]. Not documented is not the same as not supported.[1] |
| Twitter Cards | Not documented | Not mentioned for Meta AI on the Meta crawlers page[1] or the Meta AI help pages[4, 5]. Not documented is not the same as not supported.[1] |
“Not documented” means the pages read do not say; it does not mean the format is ignored.
Rich resultsThe research records no rich result type for Meta AI.
How does Meta AI crawl the web?
The research lists 5 crawler tokens for Meta AI: the names a site uses in robots.txt rules, with what each one does where the source says.
Meta-WebIndexerCrawls the web to improve the quality of Meta AI search results. Meta says allowing it in robots.txt helps Meta cite and link to a site's content in Meta AI responses. The longer form of the user agent adds a documentation link, which came back as a relative path and is not recorded here. robots.txt is presented as the control, with no statement that the crawler follows it.[1]
Meta-ExternalAgentCrawls for uses such as training foundation AI models or improving products by indexing content directly. The page shows a robots.txt rule that disallows it. The longer form of the user agent adds a documentation link, which came back as a relative path and is not recorded here. robots.txt is presented as the control, with no statement that the crawler follows it.[1]
Meta-ExternalAdsCrawls for advertising and other business products and services. The page shows a robots.txt rule that disallows it. The longer form of the user agent adds a documentation link, which came back as a relative path and is not recorded here. robots.txt is presented as the control, with no statement that the crawler follows it.[1]
Meta-ExternalFetcherFetches single links at a user's request, including to evaluate and improve agentic AI capabilities. Because a user starts the fetch, Meta says it may not follow robots.txt. The longer form of the user agent adds a documentation link, which came back as a relative path and is not recorded here. may ignore robots.txt (fetches a person asked for).[1]
FacebookExternalHitLink preview crawler for content shared on Meta's apps; not a Meta AI crawler. Listed because it is the only crawler the page ties to Open Graph. Meta says it might skip robots.txt for security or integrity checks. A short form of the user agent, facebookexternalhit/1.1, is also given. may ignore robots.txt (fetches a person asked for).[1]
How do I test my markup for Meta AI?
The research records no markup testing tool of its own for Meta AI.
Inject the World
Test a page for Meta AI
See what the research records for Meta AI against the structured data on your page.
What does Meta AI ask of site owners?
Guidelines
- robots.txt
- Sites block a Meta crawler with a Disallow rule for its token. Meta says it prefers robots.txt to non-standard signals such as NoAI tags. Crawlers may cache robots.txt for up to 24 hours, so a change can take a day to apply.[1]
- citation
- Allowing Meta-WebIndexer in robots.txt helps Meta cite and link to the site's content in Meta AI responses.[1]
- user fetch
- Meta-ExternalFetcher acts on a user's request, and Meta says this crawler "may bypass robots.txt rules".[1]
- open graph
- For the link preview crawler, Open Graph properties must appear within the first 1 MB of the page or they are cut off. This rule is about link sharing, not Meta AI.[1]
- access
- To let the crawlers in, a site can allow-list either the user agent strings or the crawler IP addresses; Meta calls the IP option the more secure one.[1]
- sources
- Meta AI may use sources and links from the internet in a response. A user can open Sources under a response and can report a source. The page names no search provider.[5]
- search partner
- A post dated 2023-09-27 said Meta AI gets real-time information in text chats through a search partnership with Bing. No later Meta page read repeats or withdraws it.[2]
- availability
- A post dated 2024-04-18 lists the United States plus Australia, Canada, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia and Zimbabwe for the English rollout. It names no search provider.[3]
What could not be verified?
These claims about Meta AI came up in the research, but no page that could be read confirms them. They are kept apart, and nothing above rests on them.
Which search provider or index supplies Meta AI's web results today, and whether the 2023 Bing partnership still holds.
Why it is open: No current primary statement. The dated 2023 statement is recorded as dated, not as current fact.
The countries where Meta AI is available today.
Why it is open: The countries field is the April 2024 list and is likely out of date.
How it was checked
Read the April 2024 post[3], which lists the United States plus thirteen other countries for the English rollout. No current availability page was collected.
Meta AI reads schema.org or other structured data from pages.
Why it is open: Not documented either way.
Sources
The pages the research used for Meta AI, numbered as cited above. Accessed November 10, 2026 unless shown.
- 1.Meta Web Crawlers (also read at the path /documentation/sharing/webmasters/web-crawlers.md on the same host) · Meta
- 2.Introducing New AI Experiences Across Our Family of Apps and Devices (published 2023-09-27) · Meta · updated January 10, 2024
- 3.Meet Your New Assistant: Meta AI, Built With Llama 3 (published 2024-04-18) · Meta · updated July 3, 2024
- 4.Meta AI and Vibes | Meta Help Center | Meta · Meta
- 5.Report a source or link used in a response by Meta AI | Meta Help Center | Meta · Meta
Market share figures: StatCounter Global Stats, AI chatbot market share, October 2026, licensed CC BY-SA 3.0.