llms.txt will not get you cited. Here is the data.
updated 2 min read Augustyn Głowacki
I found no measurement showing that llms.txt improves AI citations. In Ahrefs' May 2026 server logs for 137,210 domains, 28% published an llms.txt and 97% of those files received no requests that month (Ahrefs, 15 June 2026). SE Ranking found no citation correlation across roughly 300,000 domains (SE Ranking, 7 November 2025). Neither study tested how an AI answer changes when the file is fetched. Google says its Search features do not use llms.txt.
What does Google actually say about llms.txt?
Google says its Search features do not use extra AI text files such as llms.txt (Google AI optimization guide).
What did John Mueller say about it?
Google’s John Mueller called llms.txt “not done for search” and a “temporary crutch, perhaps to save some tokens” for AI coding tools that parse developer documentation (quoted in Ahrefs’ study and covered by Search Engine Journal). Chrome’s Lighthouse team has an experimental llms.txt check in its Agentic Browsing audits. The OpenAI, Anthropic and Perplexity documentation I checked does not name llms.txt as an input for retrieval or citations.
Who is actually requesting these files, when anyone does?
Of the 3% of llms.txt files fetched in Ahrefs’ study, bots made 96% of requests. The study classified 77% of those bots as tools other than AI, and Slackbot fetched the file more often than PerplexityBot. Tools studying llms.txt files accounted for another 12% of fetches. Ahrefs saw no AI bot request a missing llms.txt file.
So what is llms.txt actually for?
Coding agents and IDE tools can use llms.txt as an index to developer documentation, as Mueller described. Perplexity’s documentation site publishes one. The studies above measured crawler requests, so they do not show whether that index helps a coding agent.
If your documentation already lives in repository files, you can generate llms.txt during the build to give a coding agent an index. Skip it if you would have to maintain a second URL list by hand. The available measurements show no citation benefit.
What should you check instead?
Check whether search and retrieval bots can reach your pages through robots.txt, your CDN and your WAF. These include OAI-SearchBot, Claude-SearchBot, Claude-User and PerplexityBot. GPTBot serves a different purpose: model training. Check robots.txt before spending time on llms.txt.