Few three letter file names have caused as much confusion in the search world this year as llms.txt. The idea is simple enough on the surface. Site owners create a plain text file that hands large language models a tidy, curated summary of what their site contains, in theory making it easier for AI systems to find and use the right information. The harder question has been whether any of it actually matters, and Google just gave its clearest answer yet.
In an update to its guidance in June 2026, Google softened the blunt position it had taken earlier. The short version is this. Google Search does not use llms.txt, the file will neither help nor hurt your rankings, and you are perfectly welcome to maintain one anyway for other AI systems that do read it. That is less a reversal than a clarification, but it is a meaningful one, because the original wording had left a lot of people with the wrong impression.
What llms.txt is supposed to do
The concept borrows its spirit from robots.txt, the long standing file that tells crawlers where they can and cannot go. Where robots.txt manages access, llms.txt is meant to offer guidance, a structured map that points language models toward the most important pages and explains, in clean text, what a site is about. The pitch is that AI systems, which increasingly summarize and cite web content, could lean on this file to understand a site faster and represent it more accurately.
It is an appealing idea for anyone worried about how their brand shows up in AI generated answers. The catch is that an idea being appealing does not mean the major platforms have agreed to honor it. A file only works if the systems you care about actually read it, and that is exactly where Google drew its line.
What Google actually said
Google's earlier guidance discouraged the use of llms.txt and said special markup is not needed to appear in its generative AI search. That was true as far as Google was concerned, but the phrasing was broad enough to read as a verdict on llms.txt everywhere, which it was never meant to be. The revised language fixes that. Google now states that it is completely fine to create and maintain llms.txt files, or other similar files, for services or systems that use them.
Read carefully, the message has two parts. First, do not expect llms.txt to do anything for you inside Google Search, because Google's crawlers do not consume it and it carries no ranking weight one way or the other. Second, Google is not telling you to avoid it, because plenty of optimization work now aims at AI surfaces that have nothing to do with Google. The update is Google narrowing its claim to its own product rather than passing judgment on the wider ecosystem.
The keywords meta tag comparison
For anyone who has been around search long enough, llms.txt rhymes with an old ghost, the keywords meta tag. That tag let site owners list the terms they wanted to rank for, it felt productive to fill in, and search engines eventually stopped paying attention to it entirely. The worry is that llms.txt could follow the same arc, a file that feels like diligent optimization while the platforms that matter quietly ignore it.
The comparison is not perfect, since the AI landscape is younger and far less settled than search was when the keywords tag faded. But the underlying lesson holds. A signal is only as valuable as the systems willing to act on it, and self declared information that cannot be verified tends to get discounted over time. Treating llms.txt as a guaranteed lever would be repeating an old mistake.
So should you bother
The honest answer is that it depends on who you are trying to reach. If your goal is Google Search visibility, llms.txt does nothing, and your effort is better spent on the fundamentals that Google does reward, namely clear structure, strong content, fast pages, and accurate structured data. None of that has changed.
If your goal is broader AI visibility across the growing field of answer engines, maintaining an llms.txt file is low risk and may help with the systems that choose to read it. It costs little to publish one, it will not damage your standing with Google, and it keeps you ready if more platforms decide to support the format. The key is to treat it as an optional experiment for non Google surfaces, not as a shortcut to ranking.
The takeaway
Google did not change its mind so much as say what it meant more carefully. llms.txt remains irrelevant to Google Search, useful only to whatever other AI systems decide to use it, and entirely your choice to adopt. For SEOs, the practical guidance is calm and familiar. Build for humans and search the way you always have, experiment with AI specific files if you have audiences beyond Google to court, and do not mistake a tidy text file for a ranking strategy. The fundamentals still win, and they always have.
