What is llms.txt? The file for AI assistants explained
Alongside robots.txt and sitemap.xml, a third file has appeared since 2024: llms.txt. It is meant to help AI systems understand a website. What it actually achieves is less clear than some offers promise.
October 16, 2026 · 6 min read
In short: llms.txt is a text file in a website's root directory that gives AI systems an organised overview: who you are, what you offer, which pages matter. It is a proposal, not a standard, and whether ChatGPT and others use it cannot be checked from outside. Google says it does not need it. Effort and risk are low, provided you maintain it.
How AI assistants and Google's AI features deal with local businesses in general is covered in our article local SEO and AI visibility. This one is about the file itself. As of October 2026.
Where llms.txt comes from
The proposal was published by computer scientist Jeremy Howard on 3 September 2024 at llmstxt.org. The idea: websites are built for people, with navigation, ads and scripts. For a language model that is tedious to read. A short file in Markdown, a simple text format, is meant to give it a clean entry point.
The page has been revised since, most recently in August 2026, and is explicitly open to community input. llms.txt is not an official standard from any standards body, and nobody is obliged to read the file.
What goes in the file
The proposal sets out a simple order. Only the first line is required:
- A heading with the name of the website or company. This is the only required part.
- A short summary as a quote block, that is a line starting with “>”.
- Optional further text with background, without headings of its own.
- Sections with link lists: one subheading each, with links to important pages below, optionally with a short note.
- An “Optional” section for links a program can skip when it is short on space.
For a small business it might look like this:
# Muster Joinery Ltd > Joinery in Mödling, founded 1998. Made-to-measure kitchens, > built-in furniture and interior doors for private and business customers. ## Services - [Made-to-measure kitchens](https://www.example.at/kitchens.html): planning, production, installation - [Built-in furniture](https://www.example.at/built-in-furniture.html) ## Contact - [Contact and directions](https://www.example.at/contact.html): address, phone, opening hours ## Optional - [About us](https://www.example.at/about.html)
The file normally lives at /llms.txt, for example www.example.at/llms.txt. According to the proposal it can also sit in a subdirectory and then covers the pages below it. The proposal also suggests offering important pages as plain Markdown versions. For most company websites that is a second step you don't need.
The difference from robots.txt and sitemap.xml
- robots.txt controls which areas automated programs may access. It is an access rule.
- sitemap.xml lists all pages for search engines, machine-readable and usually complete.
- llms.txt is a selection with explanations. According to the proposal it is mainly read when an AI system is about to answer a question, and the sitemap is no substitute because it is too large and contains no explanations.
The three files do not replace each other. If you create an llms.txt, keep robots.txt and sitemap.xml.
What the file does and what it doesn't
Honesty pays here, because llms.txt is often sold as a shortcut into AI answers. In its documentation on AI features in Search, Google states explicitly that you don't need to create new machine-readable files, AI text files or markup to appear there (Google Search Central: AI features and your website). So for Google's AI Overviews and AI Mode the file is not needed.
Whether ChatGPT, Claude, Perplexity or other assistants fetch the file when answering a question is not disclosed in detail, and it cannot be checked from outside. llms.txt is common mainly in technical documentation, for example at Anthropic or Cloudflare. There the files help coding assistants find the right pages.
A measurable effect on enquiries has not been shown for company websites. Anyone who promises you more visibility in ChatGPT through an llms.txt is promising something nobody can deliver.
Our own llms.txt as an example
We have set up an llms.txt for koeck-media.at. It starts with the company name and a summary in German and English: location, year founded, the three service areas. Then come the company details that are also in the legal notice, and one link list each for the German and the English website. Important pages carry a short note on what they cover.
The real work is maintenance. Every new page and every new blog article goes into the llms.txt in the same round as into the sitemap. When something changes in our services, we update the description. Whether the file brings us enquiries, we cannot measure. We keep it because it costs little and gathers the key facts in one place.
How to create an llms.txt
- Create a text file named
llms.txt, saved as UTF-8 so that special characters come out right. - Write “#” and the company name in the first line.
- Below it, a summary starting with “>”: what, where, for whom. Two or three sentences are enough.
- Create sections with “##”, for example services and contact, and link the most important pages below them.
- Upload the file to the website's root directory and check it in the browser at
/llms.txt. - Set a fixed point for maintenance, for example every time a page is added.
With website builders or a CMS it is not always possible to place a file in the root directory. Some systems offer an extension for it. If it is not possible, you lose little as far as is known today.
Our view
llms.txt is a small, sensible building block, but not a lever. If you are short on time, it does not belong at the start. More important are a complete Google Business Profile, a page of its own for each service and company details that read the same everywhere. Those are the things search engines and AI assistants actually learn from. The llms.txt only summarises them. Where it fits in the bigger picture is shown in our overview marketing for small businesses.
Checklist
- Does the first line contain the company name, with a short summary below?
- Are the most important service pages and the contact page linked?
- Do name, address and phone number match the legal notice and the Google profile?
- Do all links work?
- Is the file reachable at
/llms.txtand saved as UTF-8? - Who maintains it, and when?
Frequently asked questions
What is llms.txt?
llms.txt is a text file in Markdown format that sits in the root directory of a website. It gives AI systems a short, organised overview: who runs the website, what it is about and which pages matter. It was proposed by Jeremy Howard in September 2024. It is not an official standard.
Do you need llms.txt to appear in Google's AI answers?
No. In its documentation on AI features in Search, Google states that you don't need to create new machine-readable files, AI text files or special markup to appear there. The usual SEO principles apply.
Can an llms.txt do harm?
Usually not. The file is only read when a program requests it and changes nothing on the website itself. The effort is small. What matters is keeping it up to date: a file with outdated details or dead links helps no one.
How do AI assistants see your business?
We check your presence and tell you honestly what works and what doesn't.
Get in touch