Markdown export & llms.txt

Every Velu site publishes an AI-readable copy of its content alongside the HTML. AI tools and agents can fetch the raw Markdown of any page, a site index, or the entire site as plain text — so they index and quote your docs accurately instead of scraping rendered HTML. Nothing to configure; it's generated for you.

Markdown twins

Append .md to any page URL to get its raw Markdown:

https://docs.acme.com/quickstart      → rendered HTML
https://docs.acme.com/quickstart.md   → raw Markdown

The same Markdown is returned when a request sends an Accept: text/markdown header, so a tool can ask for the Markdown version of the canonical URL without rewriting the path:

curl -H "Accept: text/markdown" https://docs.acme.com/quickstart

Site index: llms.txt

/llms.txt is an index of your site written for LLMs — the site name and a list of every page with its title, URL, and description:

https://docs.acme.com/llms.txt

An agent reads this first to discover what pages exist and which ones are relevant, then fetches the individual Markdown twins it needs.

Full content: llms-full.txt

/llms-full.txt is every page's Markdown concatenated into a single document — the whole site in one request:

https://docs.acme.com/llms-full.txt

This suits tools that want to load your complete docs into a context window at once, rather than crawling page by page.

Discovery headers

Responses advertise the index so tools can find it without guessing the path. Every page includes a Link header with rel="llms-txt" and an X-Llms-Txt header, both pointing at /llms.txt:

Link: <https://docs.acme.com/llms.txt>; rel="llms-txt"
X-Llms-Txt: https://docs.acme.com/llms.txt

Works in local preview

All of these are available in velu dev too, so you can check them before publishing:

http://localhost:8358/quickstart.md
http://localhost:8358/llms.txt
http://localhost:8358/llms-full.txt