Bingbot

Bingbot is Microsoft's main crawler for Bing Search. It fetches public pages so Bing can analyze them and decide whether to add or refresh them in its web index. Discovery, crawling, indexing, and ranking are separate decisions, so a server-log hit proves only the fetch.

The Bing index also has a documented AI retrieval role. Microsoft's Grounding with Bing Search lets AI agents search the public web, and Grounding with Bing Custom Search can restrict retrieval to selected Bing-indexed domains. A page available to Bingbot may therefore become eligible for a grounded answer and source citation, but neither crawling nor indexing guarantees selection.

Grounding is different from model training. Microsoft describes current search results being supplied to an agent for a response. A Bingbot request does not show that the page entered a foundation-model training dataset, and a later citation should be verified from the answer itself rather than inferred from crawler traffic.

Cloudflare has observed versioned forms such as bingbot/2.0 and bingbot/2.1; Bingbot is the stable token stored here. Both this bot and Cloudflare say it follows robots.txt. The fact record has no verified IP identity or request-signature directory, so the user-agent header alone remains spoofable.