llms.txt is a map, not magic
llms.txt can document selected public pages, but Google does not use it for search visibility or rankings. Other systems may ignore it, and it cannot control access.
Founder of TomHow, based in Stevenage. SEO and AI search consultancy.
Does llms.txt improve Google or AI search visibility?
Google says it ignores llms.txt for visibility and rankings, including its generative search features. The file can be maintained as optional documentation for services that choose to use it, but adoption or a citation benefit should not be assumed. Listing a page does not make it eligible for search, and an avoid list is advisory rather than an access control. Useful public pages, normal internal links and the correct crawler settings deserve attention first.
An optional public-page map for systems that choose to read it. Google ignores llms.txt for Search visibility and rankings; it is not crawler access control.
Understand what the file is for
An llms.txt file is a short public description with selected links. It can tell a reader or a system that chooses to consult it which pages the publisher considers useful. That describes an intended use, not proof that a particular answer platform reads it.
Treat it as documentation. If maintaining a map helps you keep public page facts and links accurate, that is a practical use. It should not displace work on the pages customers can actually read.
Google does not use it as a ranking signal
Google explicitly says llms.txt neither helps nor harms visibility or rankings in Google Search. That includes its generative AI features. A missing file is therefore not a diagnosis of poor Google or AI Overview performance.
For other services, check their current published guidance before claiming adoption. This note provides no measured evidence that adding the file changes retrieval, recommendations or citations on any platform.
An avoid list is not an access control
Text asking a system not to use a page is advisory. It is not a substitute for the appropriate robots.txt, indexing, authentication or server controls. Confidential information should not be made public in the first place.
Crawler roles also differ. OpenAI documents separate search and training settings; a setting for one purpose does not automatically control the other. Confirm the exact control before changing access.
Keep a public map accurate if you choose to maintain one
List only safe public pages and check that the links resolve to the intended destinations. Use the same business name and service facts as the website, and label any sample material clearly.
Review the map when a listed page changes or disappears. The file need not reproduce every sitemap URL. Normal navigation and internal links should still let a customer reach the important pages without consulting it.
Check the platform guidance.
Continue with the useful next step.
Thomas Howcroft
Founder of TomHow and senior operator for SEO, GEO, AEO, and AI search evidence.
View LinkedIn profile ->