The free report says whether AI agents can use belcanto.pt (A) today. The owner plan keeps watching: every month ChatGPT, Claude and Gemini are asked again where to find, book and order from Belcanto, and you get what changed, in plain sentences, with the score, the rank and what to fix next. Weekly rescans, an email when the score moves, the history as a trend, your menu as a page to paste, a PDF every month. No account: the link in the mail is the login.
ChatGPT, Claude and Gemini are asked where to find, book and order from Belcanto, the way people ask, with web search on. Each answer is compared with the month before and the change is a sentence: "Claude now links to your site for 'where to book'; last month it sent people to a marketplace."
One message per change, at most one a day, to the address above, with what was fixed and what broke.
The site is rescanned every week instead of every month, so a fix shows within days, and the badge follows.
Every full scan kept and drawn as a line on the plan page, with the sentence that goes with it: up, down or flat, since when.
Your menu is a PDF today. The plan reads your menu and gives you an HTML page to paste, every item and price as text, with the MenuItem block agents read. Made again when the menu changes.
The plan page as a PDF after each round, to keep or to hand to whoever runs the site.
The index serves your direct links to agents through its registry, and your direct record shows how many followed them, at no extra cost.
The deep report (19 EUR, once) asks ten questions and compares you with the places around you; the plan asks three every month and tells you what moved. Many owners take the report first and the plan after. The deep report.
The free report and its rescans stay free; the plan adds the watching, not the score. Independent: no money from marketplaces, nothing changes in the ranking. Questions: hello@directindex.org.