River CadeConcepts
Menu

Free · read-only · nothing to sign up for

See what an AI assistant can actually read about your business.

Enter a domain. This fetches the documents an AI system looks for when someone asks it about a company, and reports what came back — with the URL and the HTTP status behind every line, so you can open the same address and see the same thing. If a check cannot complete, it says so. It never reports a document as missing because the request failed.

Public GETs of the documents an agent looks for, following your site's own robots.txt. Nothing is submitted to your site, nothing is changed, and no email is required. The check runs on rcc-agent-readiness.mikeshaner78.workers.dev — River Cade Concepts' own Cloudflare account, not a third-party service.

Enter a domain to see what an AI assistant finds when it goes looking for that business. The result appears here with the URL and status behind every line, so you can repeat any of it in a browser.

What it looks for

Four documents, and whether a machine can read you.

01

A pointer at the root

/llms.txt

The file an assistant reads first to find out where a business's own account of itself lives. Without it, anything that does not already know the /.well-known convention has nowhere obvious to start.

02

A machine-readable record

/.well-known/ard.json

A structured statement of what the business is, at a path agents and crawlers already check. Marketing copy describes a business to a person; this describes it to software.

03

A catalogue of what it does

/.well-known/business-capabilities.json

The services, stated as data with the page each claim came from. Without it, an assistant has to infer the offering from page text and will sometimes infer wrong.

04

A bounded way to ask

/.well-known/agent-interaction.json

A published contract saying what an agent may ask, and what the business will never let it do — quote, book, confirm a date, or commit anything. Most businesses have a phone number for a person and nothing for software.

The rules this check runs under

It would rather say nothing than say something false.

This tool makes statements about other people's websites, which is a thing worth being careful about. These constraints are in the code, not in a disclaimer, and the check publishes them with every result.

  1. 01A failed request is unknown, never missing

    “You have no llms.txt” when the request timed out is a false statement about someone's property. Unreachable sites get no score at all.

  2. 02Nothing is inferred

    Every finding names the URL fetched and the status returned. If you cannot repeat it in a browser, it is not reported.

  3. 03It follows your robots.txt

    A path your site disallows is not fetched, and is reported as unchecked rather than absent.

  4. 04It does not grade your business

    No opinion on your design, your copy, your marketing or your search rankings. It measures whether four documents exist, and nothing else.

If the check found gaps, they are the assignment — not a verdict on the business.

Want the record built and published?

Talk about the work