llms.txt generator
Publish a map of your site for language models.
Free, no signup, and it runs entirely in your browser.
Output
Add a name to start.Save as llms.txt in your site root. It is a proposal, not a standard — nothing is obliged to read it, but the cost of publishing one is a single file.
About this tool
llms.txt is a proposal, not a standard. It is a Markdown file at your root that says what your site is and lists the pages worth reading, so a model answering a question about you has a curated route in rather than whatever a crawler happened to scrape. Nothing is obliged to read it; the cost of publishing one is a single file.
- The format is specific: an H1 with the name, an optional blockquote summary, then H2 sections of Markdown links. Generators that emit arbitrary prose produce a file that reads fine and parses as nothing.
- Curate rather than dump. This is not a sitemap — twenty pages you would actually point someone at beats every URL you have.
- A note after each link, following a colon, is part of the format and is where most of the value sits. It says what the page is for, which a title often does not.
- We publish one ourselves, at /llms.txt, generated from the database rather than written by hand so it cannot go stale.
Questions
- Do AI models actually read llms.txt?
- Some crawlers fetch it; there is no guarantee any given model uses it. It is a proposal with growing adoption, and the cost of publishing one is low enough that the asymmetry favours doing it.
- How is it different from a sitemap?
- A sitemap lists every URL for crawling. llms.txt is curated and annotated, written for a reader that needs to understand your site quickly rather than index all of it.
- Where does llms.txt go?
- At the root, as /llms.txt. Some sites also publish /llms-full.txt with the complete text of key pages inlined.