What Is llms.txt? And How It Differs From an ARD File

llms.txt is a plain text file, written in Markdown, that sits at the root of a website, at yoursite.com/llms.txt. It gives an AI model a short summary of the site and links to its most useful pages, so the model can read them without wading through menus and code.

It was proposed in September 2024 by Jeremy Howard of Answer.AI, originally to help AI coding tools read software documentation. A typical file has three parts:

• A heading with the name of the site
• A one-paragraph summary of what the site is about
• Short lists of links to the pages worth reading, each with a line saying what is there

It is a proposal, not an official standard, and each AI company decides for itself whether to read it.

So far, it's unclear. Google has said publicly that Search doesn't use llms.txt, and as far as we can tell, none of the big AI assistants has confirmed that it reads the file when answering people.

There is one sign of interest. Google's Lighthouse audit checks for llms.txt in its experimental Agentic Browsing category, next to its newer check for ARD files. That is a best-practice check, not a ranking factor.

Our read: an llms.txt file is cheap to add and does no harm. Just don't expect it, on its own, to change how often AI mentions your business.

Both are files on your own website, written for AI rather than for people. They answer different questions.

llms.txt answers what's on this website? It helps an AI model read your content. It lives at /llms.txt.

An ARD file answers what can I use here? ARD, short for Agentic Resource Discovery, is an open standard for telling AI assistants which tools and services a business offers and where to reach them (what ARD is). The file is called ard.json and lives at /.well-known/ard.json. The services that read these files are called agent finders: an AI assistant searches one when it needs something it doesn't already know about.

For a business, the difference is what the AI can do with you. Reading your website lets it repeat what your pages say. An ARD file can point it to something it can ask directly, like an MCP server, the piece that lets an AI assistant ask your business a question and get your own answer back (what an MCP server is). One lets an AI describe you. The other lets it ask you a question and use your answer.

Neither file is a guarantee. Agent finders are still early, nobody can promise that a file gets a business found, and listed is still not the same as chosen.

1. Add an llms.txt if it's easy. A heading, a two-line summary and links to your key pages is enough. Some website platforms and plugins can create one for you.
2. Check your ARD file, if you have one. Our free ARD validator checks an ard.json file against the standard.
3. Describe what you actually do. Both files are only as useful as what's in them. Say what you offer, who it's for and where, in the words your customers use.
4. Give AI something to ask, not just something to read. With Fugentic you describe your business without code, and the platform turns it into an MCP server, hosts it and registers it from the platform. It can't take bookings, orders or payments yet. It describes, and your team takes it from there.

That is the whole first step: describe your products and services to the AI ecosystem, and start preparing for how to get and serve customers in the next shift of AI.

Is llms.txt an official standard?

No. It's a 2024 proposal that some sites have adopted. ARD has a published specification, version 0.91 from August 2026.

Does Google use llms.txt?

Google has said Search doesn't use it. Lighthouse checks for it, but only in an experimental category.

Where do the two files go?

llms.txt goes at the root of your site, /llms.txt. An ARD file goes at /.well-known/ard.json.

Is llms.txt the same as robots.txt?

No. robots.txt tells crawlers what they may visit. llms.txt points AI models to what is worth reading.