If you've read anything about getting your site into ChatGPT or Google's AI answers this year, someone has probably told you to add an llms.txt file. It's an easy pitch. You drop one Markdown file at the root of your domain and AI models supposedly learn which pages matter most. The data doesn't back that up. SEO platform SE Ranking checked nearly 300,000 domains, and it found no correlation between AI citations and LLMs.txt. Both statistical analysis and machine learning showed no effect of LLMs.txt on how often a domain is cited by LLMs. Below we cover what the file is, why it doesn't move citations, and the narrow cases where it's still worth adding.
What llms.txt actually is
llms.txt is a proposed standard, introduced by Jeremy Howard of Answer.AI in September 2024, for a Markdown file at the root of your domain, for example `yoursite.com/llms.txt`. It's a short, plain-text table of contents. You write a one-line summary of the site, then link to the pages you'd most like an AI system to read.
People often compare it to robots.txt, but the two do different jobs. Unlike robots.txt, which gives bots explicit instructions on what they can or cannot crawl, LLMs.txt does not control AI behavior—it merely suggests what content might be useful. If you want to allow or block AI crawlers, that still happens in robots.txt.
What the 300,000-domain study found
SE Ranking's study is the largest public test of the file. A note on dates: SE Ranking first published the study on 7 November 2025, and the article has since been republished with a new date, 15 October 2026. Here's what it found:
- Few sites use it. In the dataset of nearly 300,000 domains, just 10.13% had an LLMs.txt file in place. Big sites use it even less: adoption among high-traffic sites was 8.27%.
- It didn't help with citations. The researchers built a model to predict how often each domain gets cited. Their model predicted citations more accurately once the llms.txt variable was removed. In other words, the file added nothing to the prediction.
- The sites AI cites most mostly don't have one. Among the 50 most AI-cited domains, only one had an llms.txt file.
One caveat: this was a correlation study, not a controlled experiment. One reviewer points out that SE Ranking compared, across 300,000 domains, the presence of an llms.txt with the frequency of AI citations; that is a correlation, not a causal test. Another notes that the article does not name the AI systems sampled or the time window. SE Ranking also sells SEO and AI-search tools. Even so, it's the biggest dataset available, and the other evidence below points the same way.
Most crawlers never ask for the file
Even if AI systems wanted to use llms.txt, they'd first have to request it. Server logs show they rarely do. Ahrefs looked at 137,000 domains, and observed 97% of files with no fetch in the month examined. A separate log analysis found something similar: across over 500M AI bot visits in a 90-day window, only 408 targeted llms.txt directly.
There's one exception. Profound, which specialises in GEO tracking, reported that crawlers from Microsoft and OpenAI are actively fetching both llms.txt and llms-full.txt files. But a bot downloading a file doesn't mean the file affects which sources get cited. And none of the major AI search providers have confirmed it uses llms.txt as a discovery or citation signal, and that holds for OpenAI, Anthropic, and Perplexity as much as for Google.
Google's position: it doesn't use llms.txt
Google has been the most direct about this. In April 2025, Search Advocate John Mueller said "To me, it's comparable to the keywords meta tag." That's the old HTML tag where site owners listed their own keywords. Search engines stopped using that tag years ago because it was easy to manipulate and unreliable.
Google has since made this official. Its AI optimization guidance says "You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search (including its generative AI capabilities)". Mueller has also explained why a file like this can't help with discovery: even a well-written llms.txt file is still self-reported information from the site that wants to be chosen.
So the problem is built into the idea. An AI system deciding which source to trust can't rely on a file where every site describes itself in its own words, just as search engines couldn't rely on keyword tags.
Where llms.txt is still useful
The file isn't useless. It solves a different problem than the one it's marketed for. It doesn't help AI decide whether to cite you. It can help an AI agent that's already working with your site find its way around. One analysis puts it this way: the clearest use for llms.txt today is documentation for coding agents and developer tools. Mueller agrees: he sees a narrow role once an agent is already on a site.
Google itself treats the file in two different ways. Chrome's Lighthouse auditing tool now checks for one, but Lighthouse checks llms.txt in an experimental agentic-browsing audit, not for rankings or citations.
Our view:
- Add one if you publish developer docs, API references, or a product that people use through AI coding tools. It's cheap to write and useful in that setting.
- Skip it, or spend 20 minutes on it at most, if you run a blog, a small-business site, or an online store and your goal is more AI citations. It won't hurt anything. SE Ranking's 300,000-domain analysis found no positive or negative effect attributable to the file's presence. It just isn't where your effort pays off.
- Be skeptical of anyone selling llms.txt as a ranking tactic. If an agency or tool charges for "llms.txt optimization" as a way to get into AI answers, the evidence above doesn't support what they're selling.
What actually gets you cited
Put the time into what AI systems can check for themselves: your pages. Answer questions clearly near the top of the page. Use specific numbers and name your sources. Keep your HTML crawlable, and make sure robots.txt doesn't accidentally block the AI search crawlers you want to let in. For discovery, Mueller pointed back to normal HTML pages and internal links. SE Ranking's study reaches a similar conclusion. One summary puts it plainly: pages that answer questions clearly and specifically still do that work.
For a step-by-step checklist, see our guide to making your content easy for ChatGPT and Google AI to cite.
FAQ
Does llms.txt help my site show up in Google's AI Overviews or AI Mode?
No. Google says Search doesn't use llms.txt or any other special AI text file, including for its generative AI features. Its Search team has also said it has no plans to support the file. Getting into AI Overviews still depends on normal SEO and well-structured pages.
Is llms.txt the same thing as robots.txt?
No. robots.txt tells crawlers what they're allowed to access, and well-behaved bots follow it. llms.txt is only a suggested reading list with no enforcement behind it. To allow or block AI crawlers like GPTBot or ClaudeBot, edit robots.txt.
Could llms.txt become important later?
It could, if a major AI provider commits to reading it in production. So far none has, and Google has said it won't. If that changes, writing the file takes an afternoon, so there's no real cost to waiting for evidence instead of acting on hype.