Mistral · Answers
MistralAI-User
Three tokens, three jobs, three directives: this is the one that fires while somebody waits. The trigger is a question typed into Vibe, Mistral's assistant surface: "When users ask Vibe a question, it may visit a web page to help answer and include a link to the source in its response." Retrieval is one-shot and scoped to that question, and Mistral rules out the alternatives, saying it is "not used for crawling the web in any automatic fashion, nor to crawl content for generative AI training." That leaves `MistralAI-Index` to sweep ahead of demand and `MistralAI-Training` to feed datasets. Verification stops at an IP match: four prefixes in a file timestamped February 2025, no Web Bot Auth and no reverse-DNS scheme documented anywhere on the page.
- Operated by
- Mistral
- Purpose
- Answers
- robots.txt token
- MistralAI-User
- Verification
- IP verified
Fetches a page in real time to answer a question someone just asked.
How to verify MistralAI-User
Mistral publishes the IP ranges MistralAI-User crawls from, and we fetch that list on a schedule. A request claiming to be MistralAI-User can therefore be checked against 4 published ranges: one that does not match is not this crawler.
Published ranges last confirmed by us on .
User agent
Mistral publishes this user agent for MistralAI-User. Match on the MistralAI-User product token rather than the whole string: vendors revise the surrounding version and URL fragments without notice.
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-User/1.0; +https://docs.mistral.ai/robots)robots.txt for MistralAI-User
Documented as robots.txt-governed with no exemption attached: Mistral says it "uses specific `robots.txt` tags" to help webmasters "manage how their sites and content interact with AI," and that "`MistralAI-User` governs which sites these user requests can be made to." Note what that is and is not: Mistral publishes no sentence reserving the right to fetch on a user's behalf despite a `Disallow`, but it also publishes no explicit compliance pledge for this token. The absence of a carve-out is inferred from silence, not stated.
Block
User-agent: MistralAI-User
Disallow: /Allow
User-agent: MistralAI-User
Allow: /robots.txt is a request, not an enforcement mechanism. It is honoured by convention, and a crawler that ignores it is stopped at your edge, not in a text file.
What blocking MistralAI-User costs you
The inline source link is what you forfeit. Mistral names this fetcher as the thing that includes "a link to the source in its response," so a disallow takes you out of the citation slot in Vibe answers precisely when a user's question is pointed at your subject. Reasoning about the block is unusually cheap, though: with no user-triggered exemption claimed, the robots directive is the intended control and no WAF rule is needed to make it hold. It is also narrow: indexing and training answer to their own tokens, so this is a genuinely independent decision rather than the first domino in an all-or-nothing choice.