Nothing is in error. The site responds. Thursday’s deploy left a noindex behind, or the robots.txt copied from staging blocks ClaudeBot, or a page lost its title. You would find out in six weeks, in a report. Here we tell you the next morning.
Google, Bing, DuckDuckGo. ChatGPT, Perplexity, Claude, Gemini, Copilot. For each one, today’s visitor count is compared to the same weekdays over previous weeks — a Tuesday against the Tuesdays. If it halves: an alert, with the figure before and the figure after.
Googlebot, Bingbot, GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended: who is allowed to read you, according to your robots.txt. Blocking the AIs is a legitimate choice — but a choice, not an oversight.
noindex in a meta tag or an HTTP header, title, description, h1, language, structured data, Open Graph. Read again every day. What would drop you out of the index raises an alert; the rest is there on the screen.
Nothing beyond Stats: the sources are read from the referrers the tag already measures, without cookies. The home-page audit needs nothing but the domain.
No keyword rankings, no “SEO score” out of 100, no forty-page report. This module says one thing only: when your visibility changes, and why.
Watch knows whether your certificate expired the day before. Uptime knows whether the page was responding. Stats sees the drop. Same project, same domains: one sentence for what actually happened.