Getting Cited by Le Chat: What You Control

Our evidence base was checked against primary sources. It has first-party crawler pages from OpenAI, Anthropic, Perplexity, Google, Microsoft and Meta. It has nothing from Mistral.

So this page will not name a Le Chat crawler user agent. It won't say how Le Chat picks sources or whether it uses a third-party search index. Guessing would be worse than saying nothing. A wrong bot name in your robots.txt does nothing, or it blocks the wrong thing.

The move is simple. Read Mistral's own documentation for its current crawler and fetcher names, and note the date you read it. Then apply the checks below. They work the same way for every AI crawler that publishes a name.

Compare this with engines that do document their bots. The ChatGPT page and the Claude page show what a documented control looks like.

Any AI answer about your brand comes from one of four stages. Splitting them shows what you can touch.

1. Parametric. What the model knows with no web access, frozen at a training cutoff. Nobody changes this in a quarter. Our evidence base doesn't record a cutoff for Mistral's models.
2. Access. Whether a crawler may fetch and store your page. This is yours and your CDN's to control.
3. Retrieval. Whether your page enters the model's context for a given question. The engine's operator controls this, with your page's topical fit as an input.
4. Generation. Which retrieved pages the model uses and cites. The model decides.

The July 2026 critical survey of 45 GEO studies (arXiv:2607.14035) found that no reviewed technique shows a stable, cross-platform causal effect on discoverability. Nearly every published GEO effect was measured on pages already inside the model's context. For Le Chat, as for every engine, your real work sits in stage 2.

robots.txt rules apply to one user agent at a time. RFC 9309 defines rules in groups keyed to a user agent. OpenAI's crawler page states the consequence plainly: each bot setting "is independent of the others." Blocking a training bot doesn't block a search bot, and the reverse is true.

That matters for Le Chat because vendors often run several agents for different jobs. The usual pattern is a training crawler, a search or indexing crawler, and a fetcher triggered by a user's request. OpenAI documents that robots.txt rules "may not apply" to its user-triggered agent. Don't assume Mistral works the same way. Check what it documents.

Here is the shape of a per-bot rule, using OpenAI's documented names as the example:

User-agent: OAI-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

The most common real-world error is a blanket AI block copied from a template. Open your robots.txt and list every AI agent it names. Add Mistral's documented agents with the rule you actually want.

Many AI crawlers download JavaScript but don't run it. Vercel's crawler study found ChatGPT and Claude crawlers fetched JavaScript files (11.50% and 23.84% of requests) but did not execute them. Later testing, as of June 2026, confirmed the same for a wider set of non-Google crawlers.

Mistral's crawlers were not in those tests. You can't know from outside whether they render. The safe assumption is that they don't.

What to do:

• Fetch your key pages with JavaScript disabled and read the raw HTML.
• Check that your pricing, product names, plan limits and one-line description are there.
• Move anything that appears only after hydration into server-rendered or prerendered HTML.

This doesn't make Le Chat pick you. It decides whether your content exists for a non-rendering crawler at all.

robots.txt can say yes while your firewall says no. OpenAI documents this conjunction for ChatGPT search: allow the bot in robots.txt and confirm your host or CDN lets its traffic through. Bot management systems can block legitimate crawlers, often with a 403. Nothing in your dashboard tells you.

The schedule changed recently too. From 2026-09-15, Cloudflare set Training and Agent crawler categories to blocked by default on ad-displaying pages. That applies to new domains, new sites on existing accounts and existing free-tier customers. A site can lose AI access without anyone editing a file.

The check:

1. Request a real page using each AI crawler's documented user agent string.
2. Record the status code and any challenge or CAPTCHA page.
3. Confirm /robots.txt itself loads for those agents.
4. Review your CDN's AI crawler settings and verified bot rules.

The AI Crawler Checker runs this request-as-each-crawler test for the crawlers it supports. For Mistral's agents, use the strings Mistral publishes.

When engines cite other sites instead of yours, being named on those sites is the route into the answer. The evidence shows a strong association here, not a promised effect. Brand size is the biggest confounder in every correlational study.

This route doesn't depend on knowing how Le Chat retrieves. If it cites a comparison page, a directory or a review site, your presence there carries over. The same holds across engines, which makes it efficient work. See third-party mentions and AI answers and the playbook for getting listed on pages AI engines already cite.

On your own site, state specific facts in plain sentences: a price, a number, a one-line definition. A fetchable page with a clear fact gives any engine a sentence it can quote. That is the only positive experimental signal for on-page content. Keep those facts consistent everywhere they appear. Contradictions across your pricing, about and footer pages are a plausible source of wrong answers.

Skip the tactics that have no credible evidence for any engine:

• Schema or FAQ schema to get cited. Ahrefs' matched test found AI Overview citations fell 4.6% after pages added schema.
• llms.txt. Google says Search ignores it. No engine in our evidence base documents reading it.
• Chunking or AI-friendly rewriting. Google rejects both. C-SEO Bench found most rewriting methods ineffective and often harmful.
• Hidden text or prompt injection aimed at AI crawlers. This is illegitimate manipulation. It exposes you to removal and reputational damage.

To see where you stand, ask Le Chat the questions your buyers ask. Run them logged out, note whether the answer shows sources, and record the date. One run is a snapshot. Answers vary by question, by retrieval state and over time.

Whatever you find, the checks above are the only part you control. For engines with documented crawler controls, compare the Perplexity page.

What is Le Chat's crawler user agent? Our evidence base contains no Mistral crawler documentation, so this page doesn't name one. Read Mistral's own documentation for its current agent names and note the date. Vendors often run separate agents for training, search and user-triggered fetches, and each needs its own robots.txt rule. A rule for the wrong name does nothing, so copy the exact strings Mistral publishes.

If I block AI training bots, will Le Chat stop citing me? It depends on which agent you block and what Mistral documents that agent does. For OpenAI, blocking the training bot GPTBot doesn't affect ChatGPT search, which uses OAI-SearchBot. Each setting is independent. Don't assume Mistral follows the same split. Check its documentation before you write a blanket AI block into robots.txt.

Does adding schema markup help me appear in Le Chat? There's no evidence it does, for Le Chat or any other engine. Ahrefs' matched-control study found adding schema produced no major uplift on any platform it tested, and AI Overview citations fell 4.6%. Existing markup does no harm, so you don't need to remove it. It just isn't a lever for AI answers.

Can a new brand get mentioned by Le Chat? Possibly, through retrieval, if Le Chat fetches the web for a given question. A model's built-in knowledge stops at its training cutoff, so a newer brand can only appear through live sources. Make sure crawlers can reach and read your key facts, and get named on established third-party pages. See [whether a brand-new website can get cited](/questions/new-website-ai-citations).

Why does Le Chat describe my brand wrongly? Common causes are an outdated training snapshot, third-party pages with old facts, or contradictions on your own site. You can't edit the model. You can fix the inconsistencies you own, make sure the correct facts are in server-rendered HTML, and correct stale listings on sites that mention you. Then re-check the answer over several weeks.