Product documentation checked November 4, 2026. Examples are illustrative; features and availability can change.

Short answer: Probably not. Google says you don’t need AI text files or special markup to appear in its AI features. A study of nearly 300,000 domains found no link between llms.txt and AI citations, and one company’s server logs showed zero requests for it from major AI crawlers. Your time is better spent on your actual pages.

Skip this if you run a developer docs site where AI coding tools are your main readers. This is for local businesses whose customers find them by searching.

If you own a six-truck plumbing company in Frisco and your agency just put an llms.txt line item on this month’s invoice, read this before you pay it. Maybe the pitch came from an SEO newsletter or a plugin offering to make one instead. It’s always the same: AI search is the new Google, and this file is how AI finds you.

Every few months a new file gets sold as the secret door into AI search. llms.txt is this year’s. The door was always your actual pages.

It’s a tidy story. The evidence doesn’t back it up. Here are the claims you’ll hear, one at a time, against what the sources actually say.

What is llms.txt, in plain English?

It’s a proposed text file you place at the root of your site (yoursite.com/llms.txt) that lists your important pages in a simple format, meant to help AI models make sense of your content. Think of it as a hand-written table of contents for a robot.

Key word: proposed. It isn’t an official standard that the big AI companies have committed to reading. That matters for everything below.

Does Google need llms.txt to show you in AI Overviews?

The myth: without the file, Google’s AI features can’t understand your site.

The fact: Google says the opposite, in writing. Its guidance on AI features states: “You don’t need to create new machine readable files, AI text files, or markup to appear in these features” (Google Search Central, AI features and your website, accessed November 4, 2026).

That’s the company that runs AI Overviews and AI Mode, telling you directly that the file isn’t required.

Do AI crawlers actually read llms.txt?

The myth: ChatGPT, Claude and Perplexity check your llms.txt every time they visit.

The fact: in one company’s server logs, they didn’t check it at all. GetMind reviewed about 218,000 requests to its own site from September 25 to October 22, 2026 and found not a single request for /llms.txt from GPTBot, ClaudeBot, PerplexityBot, Googlebot or bingbot over those 27 days (GetMind, October 22, 2026).

Be fair about what that is: one site’s logs, not a census of the internet. But if the file were central to how these crawlers work, you’d expect at least one visit in 27 days. If you’re that Frisco plumber, the file on your invoice may never be opened by the bots it’s supposedly for.

You can run the same check on your own site. Ask your host or developer to search your access logs for “llms.txt” and see who, if anyone, asked for it.

Does having llms.txt get you cited more often?

The claim here is that sites with the file get quoted by AI tools more than sites without it. SE Ranking checked nearly 300,000 domains. About 10.13% had an llms.txt file, and the study found no correlation between having one and being cited by AI tools (SE Ranking, 2025, updated 2026).

Correlation studies can’t prove much on their own. Still, if the file were a strong lever, a sample that size is where you’d see it.

If AI cites your site once, will it keep citing you?

You’ll also hear that once AI cites you, you’ve “ranked.” AI citations move around too much for that. Ahrefs tracked about 37,500 cited websites across five AI tools (AI Mode, ChatGPT, AI Overviews, Gemini and Copilot) for 20 days. Between 50% and 60% of citations were gone by the next answer that included citations, but about 75% of those came back within the 20 days (Ahrefs, October 29, 2026).

So a screenshot of your business in an AI answer proves less than it looks like, and so does a screenshot of a competitor. If you track AI visibility at all, look at the trend over a month, not a single day.

So what does AI search actually read?

Your pages. The same public, crawlable pages a person reads. That leads to some unglamorous advice, and none of it requires a new file:

Call it the same-page rule: the machine reads the same page your customer reads.

Is there any reason to make one anyway?

A small one. As far as the evidence shows, it’s harmless. If your hosting company or an SEO plugin already generated one automatically, open it and read it. Auto-generated files can list old pages or pages you’d rather not feature. Fix it or delete it, and cap the job at about 30 minutes.

What I wouldn’t do is pay for it as a growth project. At an assumed agency rate of $150 an hour, a two-hour task comes to $300 for a file the crawlers in that log study never asked for. If that’s the line item on your invoice, ask to move the $300 to one page that answers a question your customers actually ask.

Every time a new “AI file” shows up, I run it through the same-page rule. So far, none has beaten it.

Frequently asked questions

Will an llms.txt file hurt my site?

No evidence says it will. The risk is an auto-generated file that points to outdated or private pages, so read it if you have one.

Is llms.txt the same as robots.txt?

No. robots.txt is a long-established file that tells crawlers what they may and may not access. llms.txt is a newer proposal that lists content you’d like AI models to read. Blocking or allowing AI crawlers still happens in robots.txt.

What should I check instead?

Make sure your key pages load without a login, answer the main question near the top and keep important facts in plain text. Then check the trend in AI visibility monthly, not daily.