Your site, briefed for the assistants.
Every site we build serves a set of plain files written for machines: a summary of you, the whole site as text, a briefing for each major assistant, one graph of its structured data, and a robots file that lets AI crawlers in on purpose.
The files an assistant looks for.
Crawlers read a site differently from people. They look for a short account of who a site is about, the text without the menus in the way, and a list of the pages worth reading. Every AgentFaro site serves those as files at the root of its domain, on every plan.
llms.txt is the summary: who you are, your brokerage and licenses, the areas you work, your services, your key pages and your tools. llms-full.txt is the text of every page on your site in one file, each with its title and address, and none of the menus.
Then a briefing for each of ChatGPT, Claude, Gemini, Copilot and Perplexity. Each is addressed to its assistant by name and carries the same facts: who you are, where you work, what you do, how to reach you, and where to check your license.
What every site serves.
At the root of your domain, built with the site.
Checked at every build.
The files are built from the same data as your pages, every time your site is built, so they cannot describe a different agent from the one on the page. Change your bio or your areas in Agent OS, and the files change with the next rebuild.
A build that is missing any of these files fails, and so does a build where the five briefings disagree on a single fact. A failed build publishes nothing, and the site you already had stays live.
We do not count reads of these files, and a read is not a recommendation. Your website statistics show how often AI assistants read your pages, labelled as reading, and keep it apart from the visits an assistant sends you.
What agents ask about it.
What is llms.txt?
A plain text file at the root of a site that tells an AI assistant what the site is about and which pages matter, the way robots.txt tells a crawler where it may go. It is a young convention, and yours is built from your own facts.
Will the assistants read these files?
Nobody can promise that. Each assistant decides what it reads, and the conventions are new. So the same facts are also on your pages and in the structured data on every page, where crawlers already look.
Are the briefings different for each assistant?
Each one is addressed to its assistant by name, and the facts in all five are the same. That is on purpose: the whole program is about telling every machine the same story, and the build fails if they ever disagree.
Do you block any AI crawlers?
No. Your robots file lets every crawler in and names seventeen AI and search crawlers, so the welcome is explicit. Being read is where being named starts, though reading is not recommending.
Do I write any of this?
No. The files are built from the facts you gave us and keep in Agent OS: your bio, your areas, your services and your licenses. The wording around them is ours.
See what an assistant finds about you today.
The free audit checks your site and the places assistants read, in about a minute.