What is llms.txt?

Home > Glossary > llms.txt

llms.txt is a proposed website file intended to give large language models and AI tools a concise, curated index of important site content in a machine-friendly text format. It is commonly placed at the site root, while an optional llms-full.txt can provide a more extensive representation of the available material.

The proposal is conceptually similar to using a sitemap or robots.txt for machine discovery, but it serves a different purpose: it presents selected content and context in a format designed for LLM consumption. It is not a universal web standard, and support varies by AI vendor, so it should be treated as a supplementary discovery aid rather than a guaranteed ranking or citation signal.

A useful llms.txt implementation should:

• Point to canonical, high-value pages rather than duplicate every URL indiscriminately.

• Use clear headings and descriptions so a model can understand the purpose of each section.

• Stay synchronized with the live site as services, products, research, and documentation change.

• Complement, not replace, sitemaps, crawlable HTML, structured data, and sound internal linking.

For ecommerce and GEO, the practical value is organizational: the file can present the site’s most important knowledge and commercial entities in a compact form. Its impact should be tested rather than assumed, because AI systems may ignore it or use other retrieval pipelines entirely.

Example: A company can use llms.txt to point models toward its primary services, research, glossary, and documentation while keeping the canonical web pages and XML sitemaps as the authoritative public content.