Skip to content
CiteScore

Guide

How to get cited by Mistral

Le Chat is now Vibe. What Mistral publishes for web search, Open URL, and the robots.txt tokens MistralAI-User, MistralAI-Index, and MistralAI-Training. No citation rate.

20 October 2026 · CiteScore Team · 12 min read

Mistral’s chat assistant can search the web and link a source. The product page calls that assistant Vibe for work, formerly Le Chat, and says Le Chat is now Vibe. The instructions for turning web search on, and for what a sourced answer looks like, are on the Vibe Work web-search page. The crawler tokens are on Mistral’s crawler page. None of those pages publishes a citation rate, a required schema type, a word count, or an llms.txt rule. CiteScore does not log into Vibe, Le Chat, or the Mistral API. The claims below are from those three pages, read on 20 October 2026.

The same kind of split — a published control, and the files the operator does not require — is written up for ChatGPT, Perplexity, Claude, Gemini, Microsoft Copilot, Grok, Meta AI, DeepSeek, and Brave Search, and You.com, and DuckDuckGo, and Perplexity Comet. The short map is how to get cited by AI answer engines.

A source link appears when someone turns web search on

The Vibe product page says the free plan can search the web. It does not publish a daily cap. The Work docs say web search is a tool: click the + icon or type /, choose Tools, and enable Web search. A publisher cannot flip that switch. When the person asks a question that needs current information, Work searches the web and weaves the findings into the response.

The same page says a web-sourced response shows three things: a globe icon, inline links back to the original sources, and a Sources button that lists every reference. That is the documented shape of a mention. The page does not say every answer includes one. News queries are a narrower set. Work draws those from Agence France-Presse and the Associated Press. The page does not say a site outside those two agencies is dropped from other queries.

Deep Research is a separate skill in Work for a multi-source cited report. It is a feature the person chooses. It is not a publisher program, and the page does not say how a site is selected for it.

What the fetch can and cannot read

Open URL is the other documented fetch. The person pastes a URL, with Web search enabled, and Work reads that page. The limitations on that page are the closest description Mistral publishes of what a fetch keeps:

A fact that lives only on a page this URL links to, only behind a login, or only inside a widget that never becomes text, is a fact those limits say can be missed. Mistral does not say whether the fetcher runs JavaScript. The note about highly interactive sites is the published clue. Put the sentence that answers the question in the title and the first paragraph of the HTML the server sends. Mistral does not publish a word count, and this page does not invent one.

Three robots.txt tokens

The crawler page says Mistral uses specific robots.txt tags so webmasters can manage how sites interact with its AI. It names three user agents. It does not include a sample Allow or Disallow block. The token strings below are the ones that page publishes.

TokenWhat the crawler page says it does
MistralAI-UserUser actions in Vibe. When someone asks Vibe a question, it may visit a web page to help answer and include a link to the source. The token governs which sites those requests can be made to. It is not an automatic crawl, and it is not used to crawl content for generative AI training.
MistralAI-IndexAutomated crawling for indexing only. It indexes content for Mistral search, which helps answer user questions in Vibe. Content it crawls is not used for generative AI training.
MistralAI-TrainingCrawls web content to help build datasets for training Mistral models. Webmasters can disallow this user agent in robots.txt. It is not used for search indexing or to answer live user queries in Vibe.

The Training section is the only one that says webmasters can disallow the user agent. A disallow there is the opt-out that page describes, and that section says the crawler is not the one answering live queries. The User section says the token governs which sites user requests can reach. It does not say, as OpenAI says about ChatGPT-User, that robots.txt may not apply. The Index section names the indexing crawler and does not repeat the disallow sentence. This page does not invent what a Disallow does to MistralAI-Index. If you write a group, the token to write is the one in the table, not a name from a third-party crawler list.

Each token has a full user-agent string on the crawler page. The string opens a parenthesis after AppleWebKit/537.36 and does not close it. A log line can be compared with these. A client can still send a string it does not own, so a matching user agent is not proof of Mistral.

The User and Index sections link to published IP lists: mistralai-user-ips.json and mistralai-index-ips.json. Read on 20 October 2026, each file is a JSON object with creationTime and a prefixes array of ipv4Prefix values. The user file’s creationTime was 2025-02-19 and it listed four prefixes. The index file’s creationTime was 2026-04-19 and it listed two. The Training section of the crawler page does not link an IP list. This page does not paste the prefixes. Use the files Mistral hosts if you are matching an address, and treat a later creationTime as the newer list.

The wider set of tokens other operators publish, including the MistralAI-User row already in that table, is the AI crawler user-agent reference. How a disallow is written for the tokens that page covers is the robots.txt guide.

Files these pages do not ask for

Nothing on the product page, the web-search page, or the crawler page asks for a schema type or an llms.txt file. A non-empty /llms.txt is a reading list some other clients look for. The llms.txt generator and the llms.txt best practices page describe that file. They are not a Mistral requirement, and a missing file is not a penalty these three pages describe.

JSON-LD repeats facts a person can already read. It does not make Vibe cite the page. A product or an app that already shows reviews can use the Review schema generator. A service page can use the Service schema generator. A course page uses the Course schema generator. A job page uses the Job posting schema generator. A recipe page uses the Recipe schema generator. A forum or support page where people can submit answers uses the QAPage schema generator. A page that names a short passage for text-to-speech uses the Speakable schema generator. An image with a visible credit, caption, or license uses the ImageObject schema generator. A book page uses the Book schema generator. A show page or an episode page uses the Podcast schema generator. A film page uses the Movie schema generator. A dataset page uses the Dataset schema generator. A vehicle page uses the Vehicle schema generator. A hotel page uses the Hotel schema generator. A restaurant page uses the Restaurant schema generator. A clinic page uses the Medical clinic schema generator. A law office page uses the Attorney schema generator. A dental office page uses the Dentist schema generator. An agent office page uses the Real estate agent schema generator. A plumbing business page uses the Plumber schema generator. An electrical contractor page uses the Electrician schema generator. The homepage can use the WebSite schema generator. The field list is the schema guide. None of those scripts is a Mistral switch.

Free AI Answer Snapshot, then the Fix Pack

The free AI Answer Snapshot asks logged-out ChatGPT and Perplexity who they name for a local business and sends one page. AI answers vary run to run. No payment. The Fix Pack, $49 is the done-for-you work after you have seen who was named.

The free check reports GPTBot, PerplexityBot, and Google-Extended. It does not look for MistralAI-User, MistralAI-Index, or MistralAI-Training. A clear row on that check is not a Vibe citation. The GEO audit scores one public URL from the HTML we fetch and returns about ten fixes. It does not open Vibe or Le Chat.

Run the free check on the host. It reads the homepage, robots.txt for GPTBot, PerplexityBot, and Google-Extended, /llms.txt, a sitemap, and JSON-LD on the homepage. It does not log into an answer engine. Paid files are listed at checkout.

Questions about Mistral citations

Is Le Chat still the name of Mistral’s chat product?

Mistral’s product page, read on 20 October 2026, says Le Chat is now Vibe, and that conversations, settings, and plans carry over. The same page describes Vibe for work, formerly Le Chat, as the chat assistant, alongside Vibe for code. The web-search instructions live on the Vibe Work docs. This page does not describe a second, separate Le Chat crawler.

Does Mistral publish a citation rate, a required schema type, or an llms.txt rule?

Not on the product page, the Vibe Work web-search page, or the crawler page read for this guide. Those pages describe web search, Open URL, and three robots.txt tokens. They do not list a schema type, an llms.txt file, or a word count. JSON-LD and llms.txt can still be useful to other readers. They are not a switch on these pages.

Which robots.txt token is for answers, and which is for training?

Mistral’s crawler page names three. MistralAI-User is a user-triggered visit in Vibe and may include a link to the source. MistralAI-Index crawls for indexing and feeds Mistral search, which the page says helps answer user questions in Vibe. MistralAI-Training builds datasets for training models. The Training section says that crawler is not used for search indexing or to answer live user queries, and it is the section that says webmasters can disallow the user agent.

Does the free check look for Mistral’s tokens?

No. The free check reads GPTBot, PerplexityBot, and Google-Extended, plus the homepage, llms.txt, a sitemap, and JSON-LD on the homepage. It does not read MistralAI-User, MistralAI-Index, or MistralAI-Training, and it does not log into Vibe or Le Chat. A clear row is not a Mistral citation.

Sources

Request a free AI Answer Snapshot

Same request as the free snapshot page. AI answers vary run to run. This is a snapshot, not a ranking, and not a promise that any assistant will name the business.

All fields are required.

We email the snapshot to this address. We do not ask for passwords, cookies, or content behind a login. We do not sell emails. Privacy.

Email this request instead
citescore@mail.grokbot.com

No mail app opened? Copy this and email it to citescore@mail.grokbot.com from any inbox.

If that send fails, email this request instead, or use the copy buttons.

Sends the business name, city, service, website, and email to citescore@mail.grokbot.com through FlowForm. This page path is included. No account and no paid checkout.

Prepared by the CiteScore Team.