Generate your
llms.txt file.
A curated Markdown map of your most useful pages, formatted to the llms.txt convention. Fill in the fields, copy the output, publish it at your domain root.
What llms.txt is, and what it is not
llms.txt is a proposed convention: a Markdown file at the root of your domain that gives a language model a short, curated index of the pages you consider authoritative. It exists because crawling a site tells a model what pages are there, but nothing about which ones you would want cited.
It is worth being precise about its status. llms.txt is not a ratified standard, it is not read by every engine, and it is not a ranking factor in Google Search. Any tool that tells you publishing one will lift your rankings is selling you something. What it does is remove ambiguity for the agents that do read it — and the cost of publishing one is a few minutes.
What a good file contains
The convention is deliberately simple. A single # heading with your name, a blockquote summary, then ## sections with Markdown links and a short description of each. The value is entirely in the curation: a file listing every URL is just a worse sitemap.
Include the pages that answer a question definitively — your documentation, your methodology, your pricing, your primary guides. Leave out pagination, tag archives, duplicate landing pages, and anything you would be unhappy to see quoted back at you out of context.
An optional companion file, llms-full.txt, carries the same map with expanded descriptions for agents that want more context before fetching. If you publish both, keep them consistent.
Publishing and maintaining it
Serve the file at /llms.txt with content type text/plain. It should return 200 to an anonymous request — if it sits behind auth or a bot wall, no agent will ever read it.
Then keep it current. A curated index is a promise about your site's structure, and a stale one points agents confidently at pages that have moved. Review it whenever your primary navigation changes. If you cannot commit to maintaining it, a smaller file listing five pages you will not move beats a comprehensive one that rots.
Once it is live, check that crawlers can actually reach it with the AI crawler access checker — a llms.txt behind a blanket Disallow: / helps nobody.
Frequently asked questions
What is llms.txt?+
A plain-text file at the root of your domain that gives language models a curated map of your most useful pages, in Markdown. It is a proposed convention rather than a ratified standard, and it is read by some AI tools and agents but not yet by every engine.
Is llms.txt a ranking factor?+
No, and anyone claiming otherwise is guessing. It does not affect Google rankings. What it does is make the structure of your site explicit to any agent that reads it, which reduces the chance of an engine summarising the wrong page.
Where does the file go?+
At the root: https://yourdomain.com/llms.txt, served as text/plain. An optional companion, llms-full.txt, carries expanded descriptions for tools that want more context.
Does it replace a sitemap?+
No. A sitemap tells search engines which URLs exist; llms.txt tells a model which pages matter and what each one is for. They serve different consumers and you should publish both.
How often should I update it?+
Whenever your primary pages change. It is a curated index, not a crawl, so a stale llms.txt is worse than none — it points agents confidently at pages that have moved or no longer exist.