Guide · Mistral Le Chat

How to get Mistral Le Chat to recommend my product

Updated 2026-11-09 · kagentx

Short answer

Mistral's official crawler page lists two crawlers: MistralAI-User, which opens pages when a user asks, and MistralAI-Index, which builds the index for Mistral search. Both are managed with robots.txt, and content crawled for the index is not used for training. Allow both names, put the body text directly in the HTML, make each page answer one specific question, then ask 3 times yourself and look at the sources. We could not open the official pages directly; the content comes from search-result summaries and is marked unverified.

How does Le Chat search the web?

Once the user switches on web search in the Tools menu, Le Chat searches when a question needs current information, writes the results into its answer and marks the sources. Mistral's search docs (2026-11-09; we could not open them directly, so what follows comes from search-result summaries of the official pages) list these methods:

MethodWhat the official docs say (summary)What it means for you
Web searchThe answer has inline links and a Sources panel listing every referenceCited items are specific web pages, so each page must answer one question on its own
News questionsUses news partners such as AFP and the Associated Press (AP)Indie developers generally cannot get in this way
Open URLOpens only the page you give it; does not crawl the whole site or follow links; pages behind a login or paywall cannot be openedPut important content on pages that need no login
Deep ResearchMakes a search plan first, then browses several sources and produces a cited reportMany sources are cited; the more specific the page, the better the chance
Which search index is behind itNo official statement found (a second-hand report says Deep Research uses Brave; unverified)Do not guess; do the basics for both, see below

In the official docs the product name already appears as Vibe and Work; whether the rename has fully taken effect is unverified, so this article still says Le Chat.

What are Mistral's crawlers called, and how do I allow them?

The official crawler page (docs.mistral.ai/robots) lists two, with different uses:

CrawlerWhat it does (summary of the official page)Training
MistralAI-UserWhen a user asks a question in the chat, it may open a web page to help answer and include a source link in the answer; this is not automatic crawlingNot used to collect training data
MistralAI-IndexCrawls the web automatically, only to build the index for Mistral search that backs answers in the chatContent is not used for any generative AI training

Both crawlers are managed with robots.txt tags and each has its own full User-Agent (ending in MistralAI-User/1.0 and MistralAI-Index/1.0, with a link to the docs); the official page also lists each one's published IP addresses. We could not verify the full strings or the IP file URLs against the official page, so we do not copy them here. For a side-by-side of crawlers see AI crawler user agents: a side-by-side table.

To allow them, write one rule for each in robots.txt (how the rules take effect: How to write robots.txt for AI crawlers):

User-agent: MistralAI-Index
Allow: /

User-agent: MistralAI-User
Allow: /

The "block bots" switches in firewalls or CDNs ignore robots.txt, so check those too.

What can an indie developer do to make Le Chat more likely to mention them?

There is no official ranking rule. The basics below are not a guarantee for Le Chat:

  1. 1. Allow both crawlers and do not block the whole site

    Use the two rules above, and make sure there is no "User-agent: *" with "Disallow: /".

  2. 2. Put your text in the HTML itself

    Open URL reads the page content after opening it. Look for yourself first:

    # Does your own page answer without scripts? (replace the URL and the phrase)
    curl -s https://example.com/pricing/ | grep -c "a sentence from the page"

    An output of 0 means the body only appears after scripts run; see Do AI crawlers render JavaScript? for why.

  3. 3. One page answers one question a buyer would ask

    Answer directly in the first two or three sentences, with concrete numbers, and give the source and date for data. For how to write it see How to write content AI will quote: before and after.

  4. 4. Also get the traditional search basics right

    The official docs do not say where Mistral's search index comes from. If the second-hand report is right, Brave's index also matters; see How to get Brave Search to index and recommend my site.

  5. 5. Confirm in the logs that they came

    Look in the logs for the names of the two crawlers:

    # Did either Mistral bot visit? (nginx access log, adjust the path)
    grep -E "MistralAI-(User|Index)" /var/log/nginx/access.log | awk '{print $1, $4, $7, $9}' | tail -20

    Names can be faked; for more ways to check see How to check your server logs for AI crawlers.

How do I know whether Le Chat mentions me?

  1. 1. Write down 5 questions buyers ask

    Leave out your brand name and write it the way a buyer would.

  2. 2. Switch on web search in Le Chat and ask each question 3 times in new conversations

    Note three things: whether it mentions you, who it recommends, and which URLs appear in the Sources panel.

  3. 3. Test again with the same questions two to four weeks later

    For how to record it see How do you measure AI visibility? Four metrics and a table you can follow.

How is it different from other AI assistants?

On its official page Mistral splits "building the index" and "user-triggered" into two named crawlers that can be managed separately; this is close to the approach in How to get ChatGPT to recommend my product and How to get Perplexity to recommend my product. The difference is that the source of its search index is not made clear officially.

FAQ

How many crawlers does Mistral have?

Mistral's official crawler page lists two: MistralAI-User opens pages to answer when a user asks, and MistralAI-Index crawls pages to build the index for Mistral search. We could not open this page directly; the content comes from a search-result summary of it, so look at it yourself before relying on it.

Is content crawled by MistralAI-Index used to train models?

According to the summary of the official page, no: it is only used to build the search index, and the content is not used for any generative AI training. This also comes from a summary and the original text is unverified.

Can I allow or block them separately in robots.txt?

Yes. The official page says both crawlers are managed with robots.txt tags and each has its own name, so you can write separate rules. Whether user-initiated requests fully obey robots.txt, we have no official text to quote, so we make no guarantee.

Does Le Chat use Brave's search?

Only a second-hand blog report says Deep Research uses Brave's API. We did not find an official statement from Mistral or Brave; unverified.

Will Le Chat read my whole site?

In the summary of the official search docs, Open URL opens only the one page you give it, does not crawl the whole site or follow links, and pages behind a login or paywall cannot be opened. So put important content on separate pages that need no login.

Sources

  • Mistral AI · Mistral crawlers (2026-11-09 could not be opened directly; read through a search-result summary)
  • Mistral AI · Search the web (2026-11-09 could not be opened directly; read through a search-result summary)
  • kagentx evaluations: the data on the case study pages; each example gives its date.
kagentx

Do AI answers
mention you?

The evaluation checks whether crawlers can get in and whether the text is served directly, then asks ChatGPT, Perplexity and Claude (with web search on) your buyers' questions and sees whether you are mentioned. The Basic plan is $1, and results usually take 2–3 minutes.