If you run a small or mid-size business website, you probably don't need an llms.txt file right now. The proposal gives agents a cleaner map of your site. But Google's John Mueller said no AI system was using llms.txt in June 2025. In June 2026, he said it can't help LLMs choose one site over another.
llms.txt is a proposed map for AI agents
llmstxt.org describes llms.txt as a proposed Markdown file for websites that can live at /llms.txt at the site root or at a subpath, stay small, and link to more detailed Markdown pages. An agent can fetch those pages when needed.
The idea grew from a simple problem. Web pages are built for people, with navigation, ads, JavaScript, and other clutter that can make clean extraction harder for agents. The llmstxt.org site says concise information in one place can save tokens, time, and money.
The proposal says llms.txt is used most heavily for software documentation. It also gives broader examples. A business could outline its structure and policies, a personal site could share CV information, and a school could give access to course details. Still, the concept isn't limited to developer docs, even though usage is heaviest there.
The main pitch is better navigation after an agent arrives
llmstxt.org says agents are expected to view or search llms.txt, then find the information they need and follow the relevant links. It also says llms.txt is meant mainly for inference rather than training. Training runs could use it too.
The proposal describes a guide for agents and makes no claim that it improves rankings, so the file works more like a site guide than a growth tactic. The proposal also says llms.txt is designed to coexist with current standards. Sitemaps list pages for search engines. llms.txt offers a curated overview for LLMs.
For a business owner, the promise is simple. An AI assistant or agent may already be trying to learn about your company, products, policies, or documentation, and a clear file may help it find the right pages faster. The proposal doesn't say that llms.txt improves Google rankings, AI visibility, leads, or revenue. The source is silent on those outcomes.
Google's John Mueller has been clear about the limits
Search Engine Roundtable reported that in June 2025, John Mueller wrote on Bluesky, "FWIW no AI system currently uses llms.txt." The same article also quotes Mueller saying that consumer LLMs and chatbots fetch pages for training and grounding. But "none of them fetch the llms.txt file."
A year later, Search Engine Journal reported Mueller's June 2026 comments on Google's Search Off the Record podcast with Martin Splitt. He said the files can't be trusted to tell websites apart. His point was simple. A site can say it's the best. But an LLM still needs another way to decide which site best answers a query.
Search Engine Journal also reported that Mueller still saw one possible role for llms.txt. If someone is already on your website, some kind of automated system may find it helpful. That fits the proposal's focus on navigation.
Most small and mid-size business sites have higher-priority work
Based on these sources, our team wouldn't treat llms.txt as a must-have for a typical small business site today. The proposal explains what the file is for. The two later sources raise a problem. There is no sourced proof here that major AI systems are using llms.txt to choose which businesses to show people. It's optional for now.
These three readings also offer no evidence that llms.txt wins more visibility on its own. That goes for a local services site, a B2B company site, or a mid-market marketing site.
llms.txt may help sites with deep documentation, detailed help content, complex product pages, or workflows an agent may need to follow once it is already on the site. For a simpler marketing site, the sources don't give a strong reason to push it to the top of your web backlog.
Ask your web team for these things first
- Ask whether your main pages already explain your services, products, policies, and company in clear HTML copy. Search Engine Journal says Mueller pointed back to normal HTML pages and internal links for discovery.
- Ask whether your internal links make it easy to move from top-level pages to the details. Search Engine Journal says Mueller and Martin Splitt agreed HTML is still the foundation for crawling and discovery.
- Ask whether your team already has an accurate sitemap. Search Engine Roundtable quotes Mueller saying WordPress core creates sitemaps automatically for many sites. If your site is on another platform, this source does not say what happens there.
- If you have a large doc center, knowledge base, or support content, ask whether an llms.txt file could act as a curated guide for those pages. That use is supported by llmstxt.org and partly by Mueller's "already on your website" comment reported by Search Engine Journal.
- If your site is mostly standard marketing pages, keep llms.txt as a nice-to-have experiment. Based on the sources we opened, it does not look like a must-do item for most small business websites right now.
If you want help with the work that matters first, our team can support SEO, GEO, and AEO.
