What is llms.txt?
llms.txt is a plain-text summary file published at a website's root, written for AI models. It explains what the site is, what it offers and where the important pages live — without the HTML noise.
How is it different from robots.txt?
robots.txt grants access: it says which crawler may go where. llms.txt gives meaning: it presents a summary and the structure of the content. One opens the door; the other explains what is inside.
What goes in it?
- Title and a one-line definition: the brand name and what it does
- Company details: founding year, address, phone, e-mail, social accounts
- Known names: alternative spellings and abbreviations of the brand
- Service list: a link and a one-sentence description for each service
- Frequently asked questions: short and direct, in question–answer form
- Usage note: that the content may be used with attribution
Short or long?
Publishing two files is good practice: /llms.txt as the summary and link map, /llms-full.txt as the long version of every text. Models use whichever suits the need.
How do you keep it current?
A hand-written file goes stale within months. The right approach is to generate it automatically from the CMS, so llms.txt updates itself as services, pages and posts change. On Racoun's stack the file is rebuilt from the database on every request.
What happens after you publish it?
The file's existence alone guarantees nothing; the content still has to be clear and verifiable. After publishing, make sure AI crawlers are allowed in robots.txt, and reference the file from the homepage with <link rel="alternate" type="text/plain">.
