llms.txt in 2026: worth five minutes, not a strategy

Every few weeks someone asks me whether their website needs an llms.txt file. It’s a fair question. The idea has been everywhere since Jeremy Howard of Answer.AI proposed it in September 2024: a plain-text file at the root of your site, a little like robots.txt, that gives AI models a clean summary of who you are and links to your most important pages.

The short answer: it takes five minutes, it doesn’t hurt, and you shouldn’t expect it to change how AI assistants describe you. Here’s what the evidence says, and where your time is better spent.

What Google says

In June 2025, Google’s John Mueller put it bluntly: “no AI system currently uses llms.txt,” adding that it’s obvious if you look at your server logs.

Google has since written it down. Its guide to optimizing for generative AI features on Google Search says you don’t need to create new machine-readable files, AI text files, markup or Markdown to appear in Google Search, and that Google Search itself doesn’t use them. The same guide says structured data isn’t required for AI features either, though it still recommends it for rich results.

What the server logs say

In June 2026, Ahrefs looked at 137,210 websites that use its analytics. About 28% had a valid llms.txt file, a share Ahrefs calls an upper bound, since its customers lean technical. Of those files, 97% received no requests at all in May 2026.

The requests that did arrive came mostly from SEO audit tools, general web crawlers and technology scanners. AI bots made up roughly a fifth of them. In other words, the people most interested in your llms.txt file are other marketers’ software.

Where it does show up

• Lighthouse. Google’s open-source page audit tool added an “Agentic Browsing” category in version 13.3, in May 2026. It checks whether a site offers an llms.txt file, alongside things that matter more, such as a well-formed accessibility tree and layouts that don’t shift after loading. It’s a sign of where browsing agents are heading, not a ranking factor.
• Coding tools and agents. Some AI coding assistants and agents do fetch the file. For a software company with developer documentation, that’s useful. For a jeweler, a law firm or a hotel, it barely registers.

What actually moves AI visibility

• Let the crawlers in. If your robots.txt or firewall blocks the bots that AI assistants use to read pages, nothing else matters. My plain guide to AI crawlers explains which ones to allow.
• Put the facts in text. Who you are, where you are, what you make, roughly what it costs and how to reach you should be written on the page in HTML that loads without waiting for JavaScript.
• Say something original. Google’s guide puts a unique point of view, such as first-hand experience, at the center of content that does well. Rewriting what everyone else has said doesn’t.
• Keep your business data current. Your Google Business Profile and, if you sell products, Merchant Center are where Google takes many of its facts from.
• Use structured data for what it’s for. Organization, Product and Article markup still earn rich results in search, even if they aren’t a key to AI answers.

If you still want one

I keep an llms.txt file on this site, because it’s cheap and Lighthouse looks for it. If you add one, keep it short and true: your name, one line on what you do, five to ten of your most important links and how to contact you. Don’t paste your whole site into it, don’t stuff it with keywords and update it when pages move. A stale file is worse than none.

Where to start: before you write a single new file, check whether AI assistants can read the pages you already have. My free AI Visibility Check looks at crawler access, structured data and what assistants say about you, in about a minute.

Sources: John Mueller on Bluesky (June 2025, via Search Engine Roundtable), Google’s guide to optimizing for generative AI features on Google Search (updated July 2026), Ahrefs’ llms.txt study (June 2026) and DebugBear on Lighthouse 13.3 (May 2026).