llms.txt
Per-business and platform-wide llms.txt files, generated live on every request, single-locale by design, pointing agents at the richer endpoints.
llms.txt is the plain-text convention AI crawlers check for a machine-readable summary. Every published business has one, and so does the platform:
coreloop.so/{slug}/llms.txt # per business
coreloop.so/llms.txt # platform-wide
What’s in a business file
Business name, description, services, hours, all locations, contact information, and, the part that matters most, pointers to the live surfaces: the MCP endpoint and the CoreLoop Page. A crawler that reads the file once knows where to get current, structured answers forever after.
Generated live, cached sensibly
The file is generated from the current profile on each request. There is no background job that could lag and no stored artifact that could go stale: if the owner edited their hours a minute ago, the file already says so.
Served in one language
llms.txt is served in the business’s original content language. It’s a static, CDN-friendly pointer document; its job is to route agents to the MCP endpoint, which handles locale per request.
Your own domain
CoreLoop serves llms.txt on its own hosts. It cannot place a file on your own website’s domain, no third party can. What the website paste-block does instead is embed discovery markup in your pages that points crawlers to these files.