# llms.txt and agent-readable docs

A growing share of your customers never open your docs directly: they ask an AI assistant, and the assistant reads your docs for them. Clayo treats those readers as first class. Every docs site is agent-readable by default, with nothing to configure.

## llms.txt

Your site serves an index at `/llms.txt` following the llms.txt convention. It lists your project and every published article, grouped by section, with links to plain-markdown versions:

```
# Meridian

## Guides

- [What is Meridian?](https://meridian.clayo.site/articles/what-is-meridian.md)
- [Create your first monitor](https://meridian.clayo.site/articles/create-your-first-monitor.md)
```

Agents that support the convention discover it automatically.

## llms-full.txt

For agents that prefer one file, `/llms-full.txt` contains every published article in full, in your docs' display order, each with its canonical source URL. An assistant can load it once and answer questions about your entire product.

## Markdown alternates

Append `.md` to any article URL to get the raw markdown: `/articles/quickstart.md`. Clean input for agents, and handy for humans who want to paste your docs into a prompt.

## Sitemap and robots

Every site serves `/sitemap.xml` with last-modified dates, and a `/robots.txt` that welcomes crawlers. Search engines and AI crawlers both get a clean, canonical view of your docs.

## Watching agent traffic

Insights splits docs reads into agents and humans, so you can see how much of your readership is already AI. See [Insights: metrics and conversations](/articles/insights).

:::callout{type="info"}
All of these endpoints follow your canonical address. If you serve docs from a [custom domain](/articles/custom-domain) or a [subpath](/articles/subpath-hosting), llms.txt and the sitemap switch over automatically.
:::