Claim: llms.txt improves AI search visibility

The claim

“Add an llms.txt file and AI search engines will understand and cite your site more.”

Why the verdict

There is no platform documentation or published controlled study showing that llms.txt increases citations in AI search answers.

What llms.txt is

/llms.txt is a proposal (llmstxt.org, 2024) for a Markdown file at a site’s root. It gives a short summary of the site and links to key pages, optionally with Markdown versions. The aim is to help language models use the site at inference time.

Officially documented · llmstxt.org

It is a proposal, not a standard adopted by search engines. Some documentation sites publish one (OpenAI’s API docs do, for example), mostly to help coding assistants and agents read their documentation.

What platforms say

Google’s AI optimization guide lists llms.txt among things site owners can ignore for Google Search. Google Search does not use such files. Maintaining them for other services “will neither harm nor help your site’s visibility or rankings in Google Search, as Google Search ignores them.”

Officially documented · Google Search Central

Google added this note on June 15, 2026, to address questions from the community.

Officially documented · Google Search Central

Google’s AI features page says: “You don’t need to create new machine readable files, AI text files, or markup to appear in these features.”

Officially documented · Google Search Central

We found no statement from OpenAI, Microsoft or Perplexity that their search products read llms.txt for ranking or citation. Missing documentation does not prove non-use. It means the claim is unverified.

What measurements show

A 2026 study by DATA17 sampled one mainstream AI assistant on one product category (171 brands; 24 English prompts asked three times each, 72 queries). It checked robots.txt and llms.txt on 163 domains. Brands the assistant mentioned had deployed llms.txt less often (17%) than brands it never mentioned (27%). Only 18 brands were mentioned, so the authors read this as no measurable positive effect, and they do not call the file useless. Smaller brands keen for exposure may be likelier to deploy it.

Observed in practice · observed 2026-08 · DATA17 AI Research Lab · 解数智能 (unpublished report, not publicly available; readers can't check it)

Ahrefs checked the 137,210 domains in its Web Analytics that received traffic in May 2026. About 28% (38,000) had a valid llms.txt file, and 97% of those files received no requests that month. The 3% that were requested account for all the llms.txt traffic measured: 96% of those requests came from bots, SEO audit tools (including Ahrefs’s own crawlers) were the largest identified group, and AI retrieval bots made 1.1% of requests. AI bots made no requests for llms.txt files that did not exist.

Observed in practice · observed 2026-05 · Ahrefs

The authors judge that, for visibility in ChatGPT, Perplexity or AI Overviews, an llms.txt file is “largely decoration”. The closest thing to an intended audience in their data is coding agents, so the file may matter if customers use them or agents act on the site. They caution that the sample skews towards technical, SEO-aware sites, that a fetch does not prove the file was read or used, and that the data covers one month.

Observed in practice · observed 2026-05 · Ahrefs

SE Ranking’s October 2026 article reports an analysis of nearly 300,000 domains, 10.13% of which had an llms.txt file. Its correlation analysis and its XGBoost model both found no link between llms.txt and how often a domain is cited, and removing the variable improved the model’s accuracy on test data. The authors conclude that the file “doesn’t impact how AI systems see or cite your content today”, and say the results reflect the specific model and dataset tested. They still call adding the file a low-effort way to prepare.

Observed in practice · observed 2026-10 · SE Ranking

The DATA17 sample is one assistant and one category. Ahrefs counts requests to the file, not citations, and a request does not prove the file was used. SE Ranking correlates having the file with citations, and its authors note the factors interact. None of the three supports llms.txt as a visibility tactic.

What would change this verdict

• A platform documenting that it reads llms.txt for search or citation.
• A controlled experiment with matched sites and pre-registered citation measurements. See Does structured data change AI search citations? for a methodology template.

When it is still reasonable to publish one

As an agent-facing convenience for documentation sites, at low cost. Ahrefs’s data points the same way: the likeliest readers are coding agents. Just don’t report it as a visibility tactic.