More than half of web traffic is software now, and it reads nothing you designed. It reads robots.txt, llms.txt, the well-known files — or it reads nothing and leaves. Cleared for Agents is a fixed-price audit that reads your site the way an agent does: twenty surfaces, one observed API contract, a ranked fix list, and a re-scan receipt that proves the after.
Free scanners will hand you a number, and you should run one. A number is a diagnosis. It is not a fix, and it is not a defense of anything. This is a structured reading of your site at every surface an agent touches, with your API's actual contract derived from observed traffic — not from what your docs claim — and a ranked fix list a developer can apply the same week.
Where you want it done for you, we ship the files and re-scan. The after is a fact anyone can verify, including the agents you're trying to sell to.
✓ No invented findings. Every line links to a check you can re-run yourself.
✓ No rank promises. The buyer here is software, not an algorithm.
✓ No lock-in. The files are yours. The re-scan is the receipt.
Everything the audit produces, line by line. Nothing on this list is a promise about outcomes; each item is an artifact you can open, diff, or re-run.
llms.txt, catalog, server-cardn filesThe map. No /.well-known/api-catalog and an agent that wants your API finds a 404 where the inventory should be. It doesn't browse. It moves on.
The summary. Chrome's Lighthouse now checks llms.txt. Without it, every agent writes its own summary of your business from scraped HTML — and quotes it.
The proof. An MCP server listed in a registry with no server-card at the well-known path can't be verified by the thing about to call it.
Twelve checks, no signup, seconds. The scanner that grades you is the one that graded this page — the receipt below is ours, live.
Results open as JSON — the same output the $300 reading interprets.
Our own latest receipt: clearedforagents.pages.dev →