← All checks

llms-full.txt and Markdown pages checker

We look for a non-HTML, non-empty /llms-full.txt, and accept a Markdown copy of the scanned page (such as /index.md) as a substitute.

Why it matters

llms.txt only lists links, so an agent still has to visit each page. llms-full.txt puts the content itself in one file, which suits documentation and help centers. Without either, the agent spends many requests and tokens on HTML, and may stop before it reaches the page with the answer.

How to fix it

1. Generate /llms-full.txt by joining the Markdown of your key pages under H2 headings.
2. Serve it as text/plain or text/markdown, not as HTML.
3. Alternatively expose each page as Markdown at <path>.md, for example /index.md.
4. Rebuild the file when content changes, as part of your deploy.
5. Reference it from /llms.txt so agents find it.

How to verify

Run curl -sI https://example.com/llms-full.txt and confirm 200 with a non-HTML content type.

Check your site free

Scan a site

Related checks

• No walls
• Content without JS
• Headings
• Text/HTML ratio
• Markdown negotiation

All checks