What is llms.txt?
llms.txt is a proposed plain-text file placed at a website's root that summarises the site for large language models and links to its key pages. It is a proposal, not an adopted standard.Load more
Glossary · 02 Runllms.txt, in practice
The format was proposed in 2024, modelled loosely on robots.txt and sitemaps. The idea is that a model, or a tool acting for one, could read a short Markdown file to understand what a site contains instead of crawling every page.
The difference from robots.txt is commitment on the reading side. robots.txt is a published standard (RFC 9309) that every major crawler honours. No major AI vendor has committed to reading llms.txt, so publishing one is cheap but should not be expected to change how often a brand is cited.
Our published position, with sources, is that the file does not do anything measurable yet. The time is better spent on crawler directives, page structure and the off-site sources assistants actually cite.
How Surge45 builds it
In Run, as part of GEO, we audit what actually decides retrieval: robots.txt rules for AI crawlers, rendering, page structure and citation sources. We will publish an llms.txt if you want one, because it costs almost nothing, but it is not a line on our invoices and we do not sell it as a deliverable.
Read next
Generative engine optimisation (GEO)
Generative engine optimisation (GEO) is the practice of getting a brand cited and accurately described in AI-generated answers. There is no ranking position to win: a model retrieves a few sources and writes prose from them.
Answer engine optimisation (AEO)
Answer engine optimisation (AEO) is structuring content so it becomes the source an answer engine uses. An answer engine is any surface that returns a direct answer rather than links: AI assistants, AI Overviews, featured snippets and voice.
Model Context Protocol (MCP)
The Model Context Protocol (MCP) is an open standard, introduced by Anthropic in November 2024, for connecting AI applications to external tools and data sources, such as a CRM, database or file store, through a common interface.
