llms.txt

Definition

llms.txt is a proposed plain-text file at the site root that tells large-language-model crawlers which pages to prefer, in the spirit of robots.txt. It is not a standard Google ranking factor. It is not a switch that puts you in ChatGPT. Some products and crawlers look for it; many ignore it.

Treat it as a courtesy map: a short index of canonical, well-sourced URLs. Do not put secrets in it. Do not expect it to override robots.txt.

How it works

If you ship one, list the newsroom, the docs, the methods pages, and the about/entity page. Keep URLs live. Mention it in the newsroom footer if you want journalists to see the same map.

Worked example

A payments company adds /llms.txt pointing at /docs, /status, and /newsroom/incident-2026-03. Perplexity still cites the incident post because the post is indexed and specific. The file did not create the citation. The post did.

How it differs

robots.txt is the crawl control that major engines honor. Schema describes objects on a page. llms.txt is an unofficial hint. GEO work is still content and corroboration.

Common errors

Selling “llms.txt optimization.” Blocking the newsroom in robots.txt and listing it in llms.txt. 400-line keyword dumps.

Sources

• GEO: Generative Engine Optimization — Retrieved synthesis
• Google AI features docs — Indexed snippet eligibility
• llmstxt.org — Proposed convention