llms.txt Is an Index, Not an AEO Strategy

The useful version

An llms.txt file can be a readable directory of a site’s most important public resources. It can state the canonical identity, describe the site briefly, and link to high-value guides, data, tools, policies, and contact information.

That is cheap to maintain when it is generated from the same source as the website. It may help any consumer that deliberately looks for it.

The claims it cannot support

The file does not prove that a crawler fetched the linked pages, that a model used them in training, that a search product indexed them, or that an answer system will cite them. Google’s current AI-search guidance says no special AI file is required and specifically gives llms.txt no strategic role.

It is also not a security boundary. Private URLs remain private through authentication and origin design, not a voluntary text file.

A sane implementation

• Keep the file under version control or generate it deterministically.
• Link only canonical public resources.
• Use ordinary Markdown headings and links.
• Include a short content/license note when relevant.
• Test every link in continuous integration.
• Update it when navigation and priority resources change.
• Measure requests and downstream behavior before investing more.

This site publishes llms.txt because the maintenance cost is tiny. The strategic work remains the pages it points to.