Google's llms.txt Split: Search vs Lighthouse

What is llms.txt and what is it for?

llms.txt is a proposed text file you place at your domain root to give AI crawlers a clean, curated map of your most important content. Think of it as a robots.txt aimed at language models. The idea is to help AI systems find and understand your pages. It is a community proposal, not an official standard.

Does Google Search use llms.txt?

No. Google's May 15 generative AI search guidance states that no special AI files or markup are needed for its search systems. John Mueller confirmed it plainly, saying the file is not used for search. If your only goal is Google ranking or AI Overviews, an llms.txt file does nothing for you today.

Why does Chrome Lighthouse check for llms.txt then?

This is the confusing part. Chrome's Lighthouse 13.3, released May 20, added an Agentic Browsing audit that checks whether your site has an llms.txt file. So one Google product ignores it while another flags its absence. The split reflects different teams betting differently on where agentic browsing is headed, not a unified Google position.

What did John Mueller say about it?

On Bluesky on May 20, Mueller addressed the irony directly. He wrote that the short answer is it is not done for search, adding that there is more to websites than just SEO. The point: Lighthouse checks many things that help users or agents without being Google Search ranking factors. The two are not the same.

When does an llms.txt file actually help?

It may help with specific agentic tools and crawlers that choose to read it, mostly in documentation-heavy or developer contexts. The honest answer is that adoption and benefit are both unproven. SE Ranking found removing llms.txt from its AI-citation model improved accuracy, suggesting the file added noise, not signal, for citations today.

Which sites should bother with llms.txt?

Mostly large documentation sites, developer platforms, and tools whose users run AI agents against them. For a typical B2B SaaS marketing site, the payoff is near zero right now. If you can add it in ten minutes and maintain it, there is little downside. If it competes with real work, skip it for now.

How do AI crawlers behave toward llms.txt today?

Inconsistently. Some agentic tools look for it, most major AI crawlers do not yet rely on it, and adoption sits around 10% of sites per an SE Ranking study of 300,000 domains. There is no enforcement and no guarantee any given model reads it. Treat it as experimental infrastructure, not a dependable visibility lever.

Should a B2B SaaS add llms.txt now?

Only if it is cheap to maintain. The smarter investment is clean page structure, direct answers, and strong content, which every AI system already rewards. If you have spare time, add a simple llms.txt pointing to your key pages. But do not expect measurable AI visibility gains from it in 2026.

Can llms.txt hurt your SEO?

No, it will not hurt your Google rankings. Google ignores it for search, so its presence or absence is neutral for SEO. The only real cost is your time to create and maintain it. A stale llms.txt that points to dead or outdated pages helps no one, so keep it current or leave it off.

Will Google ever officially support it?

Unknown, and worth watching rather than betting on. Google has not committed to llms.txt for Search, and Mueller's comments suggest no near-term plan. The Lighthouse audit shows some internal interest in agentic browsing, but that is not an official ranking signal. Build for clear content now, and adopt llms.txt formally only if Google does.

Sorting out what actually moves AI visibility? Pair this with my guide to AEO, GEO, and SEO, the breakdown of why ranking #1 no longer wins citations, and the Reddit AI citation index piece. Let's chat.