In short. MistralAI-User identifies user-directed page requests in Vibe, with a published user-agent string and IP address feed for checking request origins.

Definition: MistralAI-User is the user-agent identifier for web requests prompted by user actions in Mistral's Vibe assistant. A user-agent identifier is the name a client sends to identify itself when requesting a resource.

Key facts

As of 17 October 2026, Mistral's crawler documentation assigns MistralAI-User to user actions in Vibe and excludes automatic web crawling from its role.
Mistral's published IP feed contains four IPv4 prefixes under the prefixes field.
The Robots Exclusion Protocol, RFC 9309, uses a crawler's product token to select the relevant robots.txt rules.

When does MistralAI-User request a page?

Mistral's MistralAI-User description says that a question asked in Vibe may lead the assistant to visit a web page to help answer it. The request therefore belongs to a user-directed interaction. The documentation separates this activity from automatic crawling.

For a website operator, this identifies the purpose associated with the token in a request log. Establishing that a particular request comes from Mistral also requires checking its origin against the published verification data.

How can a request be checked?

An HTTP request carries headers, which are named fields describing the request. Mistral's crawler information publishes this full user-agent value:

Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; MistralAI-User/1.0; +https://docs.mistral.ai/robots)

The documentation links directly to the MistralAI-User IP address feed. An excerpt from its prefixes array is:

{
  "ipv4Prefix": "20.240.160.161/32"
}

A practical check compares the request's source IP address with the prefixes in that feed, alongside the user-agent value. The feed supplies machine-readable address data, so an automated check can read its ipv4Prefix fields rather than maintain a manually copied list.

Checking origin addresses is an established verification method: Google's request verification documentation describes matching crawler IP addresses against published ranges and warns about clients claiming another operator's identity. For MistralAI-User, the relevant comparison data comes from Mistral's own feed.

How are site preferences expressed?

A robots.txt file expresses crawler access rules. RFC 9309 places it at /robots.txt and defines groups selected through a User-agent line. Applying that syntax to Mistral's token gives this refusal group:

User-agent: MistralAI-User
Disallow: /

The group expresses a refusal for all paths. Under the standard, matching groups are combined; the wildcard User-agent: * group applies when no specific group matches. Existing restrictions therefore need review when adding a named group. The standard also treats robots.txt as a crawling policy: private resources require access controls such as HTTP authentication.

FAQ

Should I put the full browser-like string in robots.txt?

Use MistralAI-User as the product token. RFC 9309 distinguishes the token used for rule selection from the longer identification string sent in the HTTP header.

Where can I find Mistral's verification addresses?

Use Mistral's MistralAI-User IP feed. Its prefixes array contains entries with an ipv4Prefix field, which can be compared with the source address recorded for a request.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer