What llms.txt is and whether your website needs one
llms.txt is a file that summarises your website for AI. Here is what it is, what it should contain, what it does not do and how to create yours.
You probably know robots.txt, the file that tells search engines what they may crawl. Since 2024 a similar idea aimed at artificial intelligence has been spreading: llms.txt. Besides, it is catching on fast, so it is worth knowing what it is and what it is not.
What llms.txt is
It is a text file placed at the root of your website that summarises, in a simple format, who you are, what you offer and which pages matter most, with a link and a sentence for each.
Specifically, Jeremy Howard published the proposal in 2024. That way, a language model consulting your website gets a clean map, without menus, ads or code, of what really matters.
What llms.txt is not
- Not an official standard adopted by every search engine or AI.
- No guarantee that any assistant will mention you.
- Not a replacement for your website's content, your sitemap or your robots.txt.
It is a small, cheap help, not a magic lever.
What your llms.txt should contain
For a small business, the file should include:
- The business name and a sentence saying what you do and where.
- Services with indicative prices, if you publish them.
- The important pages with a link and a short description: services, pricing, work, contact.
- Contact details and the area you serve.
Typical structure:
- A heading with the business name.
- Then a summary paragraph.
- A services and prices section.
- Finally, a list of pages with links.
- Contact details.
Mistakes to avoid
- Saying something different from the website. If you change prices, change both.
- Including everything. It is not a dump of the whole website, it is an index.
- Exaggerating. AI cross-checks, so inflated claims cost credibility.
- Forgetting it. When you add new pages, add them here too.
Is an llms.txt worth it?
On the one hand, if your website is already well built, creating the file takes half an hour and has no downside. On the other hand, if your website has deeper problems (not indexed, does not explain what you do, no prices or facts), this file will not fix them: start there.
What matters most for appearing in AI answers is still what I explain in how to get recommended by ChatGPT: clear content, structured data, reviews and mentions.
How I do it
Every Nueva Página website includes one, generated from the website's own content so it always says the same thing. For example, you can see this website's at nuevapagina.com/llms.txt.
An example for a local business
Picture a physiotherapy clinic. Its file would open with the name and one sentence: "Movement Clinic, sports and pelvic health physiotherapy in Leeds". Then the services with indicative prices, the area and opening hours. Finally, the list of important pages, each with its link and a sentence saying what is on it.
Nothing more is needed. In fact, the shorter and clearer it is, the better AI can use it. The original proposal, with its format and examples, is at llmstxt.org. And since this file only summarises what your website already says, the website itself should be well prepared first: see how to be recommended by ChatGPT. Website first; summary second.
Frequently asked questions
Where does the file go?
At the root of the domain, just like robots.txt: yourbusiness.com/llms.txt.
Do I still need a robots.txt?
Yes, because they are different things. While robots.txt says what bots may crawl, this one summarises the content.
Can I block AI from reading my website?
You can say so in robots.txt for crawlers that respect it, but if you want to be recommended, you want the opposite.
Will my website have an llms.txt?
Yes. Yes, every one includes it, generated from the website's own content. See how I work.