llms.txt
llms.txt is a proposed convention for a plain-text Markdown file at a website's root that gives AI systems a short, curated guide to the site's most useful pages.
Last updated October 3, 2026
What the proposal says
Jeremy Howard of Answer.AI proposed llms.txt in September 2024. The file sits at /llms.txt and is written in Markdown: a title, a short summary and lists of links to key pages, each with a one-line description. The idea is that a language model with limited context can read one small file and find the pages that matter, instead of parsing a full site. The proposal also describes optional companion files with full page content in Markdown.
It is a community proposal, not an internet standard. It is different from robots.txt, which tells crawlers what they may access. llms.txt does not grant or block access. It is a reading guide.
Adoption and evidence
Many documentation sites and some companies publish the file, and developer tools may read it when a user supplies a site to an assistant. As of this entry's date, no major AI provider has confirmed that its search or training systems use llms.txt for retrieval or ranking, and Google representatives have publicly said Google Search does not use it. There is no published evidence that having the file increases how often a brand is mentioned or cited. Treat it as low-cost housekeeping with an uncertain return, and check provider documentation for any change.
Example
A hypothetical example: a company lists its newsroom, leadership page, fact sheet and corrections policy in llms.txt, each with a short description. If a tool does read the file, it reaches the accurate reference pages first. The same pages should also be clear, linked and crawlable on the site itself, since that is what search indexes and assistants retrieve today.
Why it matters to communications and risk teams
The file is within a comms team's reach because it points at pages the team already owns. The sound reasoning is that clear, current owned pages matter whatever the file does; see generative engine optimization and AI citations. Measuring whether anything changed requires repeated benchmark questions before and after, and a single run shows nothing because answers vary. The free AI Perceptions check samples answers for a brand.
Common misconceptions
No provider has documented llms.txt as a ranking signal, and the file is not an official standard.
It cannot block AI crawlers, since it grants or denies no access. It also cannot correct false claims that appear on other sites.
See how these signals show up in your own coverage on the PeakMetrics platform or run the free AI Perceptions check. Back to the glossary.