The short definition
llms.txt is a machine-readable summary file published at the root of a website. Its job is to help AI agents understand what the site is, which pages matter most, and which URLs should count as canonical sources.
What it helps with
The file helps with routing. Instead of making a model infer your best product, pricing, docs, security, or contact pages from a large site, it points directly to them.
- It makes the site easier to orient quickly.
- It reduces reliance on secondary summaries or outdated pages.
- It gives teams a small place to add practical guidance.
What it does not do
llms.txt does not replace strong underlying pages. If your homepage is vague or your pricing and trust pages are weak, the file can only route an agent into weak content faster.
Who should publish one
The biggest gains usually show up on sites with meaningful product, docs, pricing, support, trust, or developer content. If there are multiple plausible pages an agent could cite, the file becomes more valuable.
Next step: Draft a file in the builder, then use the score page to see whether the site underneath it is ready to support the routing.