Does llms.txt matter? What the data says

On the evidence published so far, llms.txt does not matter for visibility in Google or in AI search. Google states that its Search does not use the file, and the two named studies available both found no measurable effect. The file is harmless and has a real use for developer documentation, but it is not an SEO task. This article sets out what the file is, what the search engines have said and what the data shows, including the limits of that data.

What llms.txt is

The llms.txt proposal was published by Jeremy Howard on 3 September 2024. It suggests placing a Markdown file at the root of a website, at /llms.txt, that gives AI systems a short, curated guide to the site: a title, a summary and lists of links to the most useful pages, ideally in clean Markdown versions.

The proposal says the file is mainly intended for use at inference time, meaning when an AI agent is helping someone and needs to read a site, and not mainly for training. It is a proposal, not a standard that any search engine has agreed to follow. The proposal's own site says thousands of sites now publish one.

What the search engines say

Google

Google's guide to optimising for generative AI features, last updated in July 2026, addresses the file by name. It says site owners do not need to create new machine-readable files, AI text files, markup or Markdown to appear in Google Search, because Google Search does not use them. It adds that creating llms.txt files will neither harm nor help a site's visibility or rankings in Google Search.

One detail has caused confusion. Chrome's Lighthouse testing tool includes an llms.txt check. Chrome's documentation for that check describes the file as optional, treats a missing file as not applicable and makes no claim about search. It concerns AI agents browsing a site, which is a separate matter from ranking.

OpenAI

OpenAI's crawler documentation and its ChatGPT search help page, as read on 10 October 2026, describe robots.txt rules and crawler access as what decides whether a site can appear in ChatGPT search. Neither page mentions llms.txt.

What the data says

Two named studies have been published. Both come from SEO software companies, and both point the same way.

SE Ranking: no link to AI citations

In November 2025, SE Ranking analysed nearly 300,000 domains. According to the study:

• 10.13% of the domains had an llms.txt file.
• Adoption was almost even across small, medium and large sites.
• There was no correlation between having the file and how often a domain was cited in AI answers.
• When the researchers removed the llms.txt variable from their prediction model, the model became more accurate, which they read as the file adding noise and not information.

The article does not say which AI systems were sampled or over what period. The authors note that the result depends on their model and dataset.

Ahrefs: most files are never requested

In June 2026, Ahrefs studied 137,210 domains using server log and analytics data for May 2026. According to Ahrefs:

• 28% of the domains published an llms.txt file. Ahrefs calls this an upper bound, because its customers are more technical than the web as a whole.
• 97% of those files received no requests at all during the month.
• Of the requests that were made, AI retrieval bots, the kind that fetch pages to answer a live question, accounted for 1.1%. AI training crawlers accounted for 5.3%, and AI agents and coding tools for 10.5%.
• SEO audit tools made more requests for the files than any category of AI system.
• No AI bot requested the file on sites that did not have one, so AI systems do not appear to go looking for it.

Ahrefs points out that a request does not prove the file was read or used, so even these small figures are ceilings. The study covers one month.

How strong is this evidence

It is thin, and it should be described that way. Only two named studies were found. Both come from companies that sell SEO tools, both were published on company blogs, and one covers a single month. One measures correlation with citations. The other measures requests in server logs. Neither can prove that the file never helps.

What can be said is narrower. Google has stated plainly that its Search does not use the file. Neither study shows a benefit. The one group of AI tools that fetches these files in any number is AI agents and coding assistants.

Where the file does help

The Ahrefs data shows real use in one place: AI agents and coding tools. A company that publishes developer documentation, an API or technical reference material has a reason to offer an llms.txt file, because coding assistants read documentation on a developer's behalf.

A hotel, clinic or local service business is not in that position.

What Stryx recommends

• Do not treat llms.txt as an SEO or AI search task. Neither study shows that it earns visibility.
• If a website platform generates one automatically, there is no need to remove it. Google says it neither harms nor helps in Search.
• If one is published, keep it accurate. It is a public file that lists the pages a business considers most important.
• Put the effort into the basics: pages that can be crawled and indexed, one clear topic per page, a short summary at the top and real questions and answers.
• Check crawler access. For ChatGPT search, robots.txt and firewall settings decide whether a site is eligible. That is covered in the guide to getting recommended by ChatGPT.

This picture may change if a major AI search provider announces support for the file. As of 10 October 2026, none of the sources read for this article shows that one has.

For a short definition, see the glossary entry on llms.txt. Stryx covers this under AI search optimisation, and the wider technical checks are in the technical SEO audit checklist. To discuss a site, get in touch.