Glossary
MistralAI-User Mistral AI's user-triggered fetcher.
The bot that reads your page when someone asks Mistral's assistant, Vibe, a question. Prefer's crawler guides cover where it sits in Mistral's three-token family, so you can block training without blocking live user requests.
MistralAI-User is Mistral AI's fetcher that visits a page when a user asks its assistant, Vibe, a question, and may link the page as a source. Prefer's guides show how to allow it in robots.txt.
Key facts
At a glance.
The entity facts an assistant lifts first, each from Mistral's own crawler page.
- Operator
- Mistral AI · mistral.ai
- Type
- User-triggered fetcher for Vibe, Mistral's assistant
- User agent token
- MistralAI-User
- Full UA string
- Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-User/1.0; +https://docs.mistral.ai/robots)
- Respects robots.txt
- Yes. Mistral says the token governs which sites user requests can be made to
- IP ranges
- Published at mistral.ai/mistralai-user-ips.json
- Used for AI training
- No. Mistral says it is not used to crawl content for generative AI training
- Affects citations
- Not documented. Fetched pages may be linked as the source in a response
- First announced
- Not documented
How it works
What happens when MistralAI-User visits.
Three steps, all triggered by a person. There is no background crawl: the visit happens because a Vibe user asked something.
- 01
A user asks Vibe
Mistral says MistralAI-User is for user actions in Vibe. It is not used for crawling the web in any automatic fashion.
- 02
Robots.txt is checked
Mistral says the MistralAI-User token governs which sites these user requests can be made to, so your robots.txt group for it decides whether the fetch can happen.
- 03
The page informs the answer
Vibe may visit the page to help answer and include a link to the source in its response. Mistral says this content is not used for generative AI training.
The Mistral crawler family
MistralAI-User is one of three. They do different jobs.
Each has its own token, so you can block training while keeping live user requests and search indexing allowed.
Purpose, what blocking it changes, and whether it feeds Vibe's answers.
| Bot | Purpose | Blocking it means | Feeds Vibe answers |
|---|---|---|---|
MistralAI-UserThis term | Fetches a page when a Vibe user asks a question, and may link it as the source. Not used for automatic crawling or training. | Vibe cannot fetch your pages on a user's behalf. | ✓ Yes |
MistralAI-Index | Crawls the web automatically to index content for Mistral search, which helps answer questions in Vibe. Not used for generative AI training. | Your pages are left out of Mistral's search index. | ✓ Yes |
MistralAI-Training | Crawls web content to build datasets for training Mistral's generative AI models. Not used for search indexing or live user queries. | Your pages are excluded from Mistral's training datasets. | ✕ No |
Why it matters for AEO
Allow or block: the decision in one sentence.
Should you block MistralAI-User?
For a brand that wants AI visibility, no, and Prefer's crawler guides give the same answer for every user-triggered fetcher. MistralAI-User only visits when a Vibe user asks a question, Mistral says it is not used for training, and the fetched page may be linked as the source, so blocking it means Vibe cannot read your page when a user asks. If training is the concern, block MistralAI-Training instead.
- You want Vibe to read and link your pages when users ask about your category
- You already allow ChatGPT-User, Claude-User or Perplexity-User
- Your concern is training, which Mistral handles with a separate token
- You do not want any AI assistant reading pages on a user's behalf
- The content is paid or licensed and should not be quoted in answers
- You block by path, so public pages stay reachable
Mistral does not document what a block does to citations. The guidance above follows from what it does document: the bot fetches for a user and may link the source.
Allow or block it
Three robots.txt patterns that cover most cases.
Each pattern names Mistral's tokens explicitly. A wildcard Disallow catches all three, which is how sites block live user requests by accident.
robots.txtCopy the pattern that matches your decision above.
User-agent: MistralAI-Training
Disallow: /
User-agent: MistralAI-User
Allow: /
User-agent: MistralAI-Index
Allow: /
The pattern most brands want: out of Mistral's training datasets, still reachable for Vibe users and Mistral search.
- Check for a wildcard firstA User-agent: * group with Disallow: / already blocks all three Mistral tokens. A group that names one overrides the wildcard for that token only.
- Three tokens, three decisionsBlocking MistralAI-Training does not block MistralAI-User or MistralAI-Index. Name each one you want to control.
- Blocking is not removalA Disallow governs future requests. It does not reach content collected before the rule.
User-agent: MistralAI-User
Allow: /
User-agent: MistralAI-Index
Allow: /
User-agent: MistralAI-Training
Allow: /
The default when no rule names Mistral, stated explicitly so a future wildcard block does not catch it.
- Check for a wildcard firstA User-agent: * group with Disallow: / already blocks all three Mistral tokens. A group that names one overrides the wildcard for that token only.
- Three tokens, three decisionsBlocking MistralAI-Training does not block MistralAI-User or MistralAI-Index. Name each one you want to control.
- Blocking is not removalA Disallow governs future requests. It does not reach content collected before the rule.
User-agent: MistralAI-User
Allow: /
Disallow: /research/
Disallow: /members/
Let Vibe users reach public pages, protect the paid archive.
- Check for a wildcard firstA User-agent: * group with Disallow: / already blocks all three Mistral tokens. A group that names one overrides the wildcard for that token only.
- Three tokens, three decisionsBlocking MistralAI-Training does not block MistralAI-User or MistralAI-Index. Name each one you want to control.
- Blocking is not removalA Disallow governs future requests. It does not reach content collected before the rule.
Verify a visit
How to confirm it was really MistralAI-User.
Any script can claim the MistralAI-User user agent. Mistral publishes the IP addresses it fetches from.
- 01
Match the user agent
Look for the MistralAI-User token in the user agent field. The full string ends with a link to docs.mistral.ai/robots.
- 02
Match the IP address
Compare the request IP against the list Mistral publishes at mistral.ai/mistralai-user-ips.json. A MistralAI-User user agent from an IP outside that list is a spoof.
203.0.113.42 - - [01/Oct/2026:09:14:07 +0000] "GET /pricing/ HTTP/1.1" 200 18422 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-User/1.0; +https://docs.mistral.ai/robots)"In context
The term in a sentence.
01"We disallowed MistralAI-Training but kept MistralAI-User and MistralAI-Index allowed, so Vibe can still read and link our guides."
02"The MistralAI-User hits in our logs line up with questions people asked Vibe, not with a scheduled crawl."
Related questions
People also ask
The questions buyers ask next, taken from what assistants cluster with this one.
Is MistralAI-User a training crawler?
No. Mistral says MistralAI-User is not used to crawl content for generative AI training, and Prefer's crawler guides class it as user-triggered. Training is handled by a separate token, MistralAI-Training.
See the crawler comparison →What is the MistralAI-User user agent string?
At the checked-on date Mistral publishes it as Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-User/1.0; +https://docs.mistral.ai/robots). Prefer's robots.txt guides use the token on its own, MistralAI-User, in the User-agent line.
What is Vibe?
Vibe is the name Mistral's current crawler docs use for its assistant; older third-party write-ups call it Le Chat. Prefer's crawler guides use Mistral's own name, and MistralAI-User is the bot that fetches pages when a Vibe user asks a question.
Questions
Asked plainly.
What does MistralAI-User do?
MistralAI-User is the fetcher Mistral AI uses for user actions in Vibe, its assistant: when a user asks a question, it may visit a web page to help answer and link to the source. Prefer's crawler guides class it as user-triggered, like ChatGPT-User and Claude-User. Mistral says it is not used for automatic crawling or to collect content for generative AI training.
How do I block or allow MistralAI-User?
Prefer's free robots.txt generator builds the groups for the major AI crawlers, and you add a 'User-agent: MistralAI-User' group to that file by hand. Use 'Allow: /' to let Vibe fetch your pages for users, or 'Disallow: /' to refuse. Mistral says the token governs which sites these user requests can be made to.
What is the difference between MistralAI-User, MistralAI-Index and MistralAI-Training?
MistralAI-User fetches a page live when a Vibe user asks a question, MistralAI-Index crawls the web to build the index for Mistral search, and MistralAI-Training collects content for training Mistral's models. Prefer's crawler guides treat them as three separate decisions, because Mistral documents a separate robots.txt token for each. Mistral says neither the User nor the Index bot is used for generative AI training.
Does blocking MistralAI-User stop Vibe from citing my site?
Mistral does not document the effect of a block on citations. Prefer's crawler guides stick to what it does say: the token governs which sites user requests can be made to, and fetched pages may be linked as the source in Vibe's response. Blocking it therefore means Vibe cannot fetch your page when a user asks.
Sources
Where these facts come from.
Every fact on this page traces to one of these. Where Mistral does not document something, the page says so rather than guessing.
- 01 Mistral AI, Mistral crawlers The documentation for MistralAI-User, MistralAI-Index and MistralAI-Training, including tokens, UA strings and purpose. Vendor docs 1 Oct 2026
- 02 Mistral AI, MistralAI-User IP list The published IP list used for the verification step above. Vendor data 1 Oct 2026
- 03 Mistral AI, MistralAI-Index IP list The published IP list for Mistral's search indexing crawler. Vendor data 1 Oct 2026
- 04 Robots Exclusion Protocol, RFC 9309 The standard that defines how user-agent tokens, Allow and Disallow rules are interpreted. Standard 1 Oct 2026
Mistral AI, Vibe and the MistralAI user agent names belong to Mistral AI. Prefer is not affiliated with, endorsed by or sponsored by Mistral AI. This entry reflects public documentation at the dates shown. Something wrong here? Tell us and we will fix it →
Keep reading