llms.txt
llms.txt is a proposed Markdown file at the root of a website that lists its most useful pages with short notes, so language models and AI agents can find clean content when answering a question.
llms.txt is a proposal published by Jeremy Howard at llmstxt.org in September 2024 and revised as version 2 in August 2026. The file sits at /llms.txt and is written in Markdown. Only an H1 with the site or project name is required; it usually adds a short summary in a blockquote and H2 sections of links, with an “Optional” section for links an agent can skip. The proposal also asks for clean Markdown copies of key pages.
It is aimed mainly at inference, the moment a tool fetches context to answer a question.
Why it matters now: it is widely discussed and often sold as a way to get cited in AI answers. Google says Google Search does not use llms.txt and that keeping one neither helps nor harms rankings. Ahrefs found that 97% of the llms.txt files on 137,210 domains it studied received no requests in May 2026.
Practical note: publish one if developers or coding agents read your documentation. For other sites it is optional and low priority.