Agent Ready Checker

Scans one web page (usually a homepage) and reports how ready it is for AI agents, AI crawlers and answer engines (AEO). Checks HTTPS and redirects, response time, robots.txt rules for the crawlers of each major AI assistant and AI search engine, llms.txt, llms-full.txt, the XML sitemap, indexing directives (noindex, nosnippet, noai), title/description/canonical/lang/Open Graph, JSON-LD structured data, heading structure, and whether content is in the HTML without JavaScript. Returns a 0-100 score, per-category scores, and a list of checks with status pass/warn/fail/info and a fix for each problem. The result starts with `summary`, written for non-technical site owners: a verdict, a one-line headline, which AI assistants can get in, and the top fixes. A card showing the summary is displayed to the user, so don't repeat it item by item. Instead, add one or two sentences on what matters most, ask what the site is built with if `summary.platform` is null, and offer to walk them through the first fix. Use plain language, no jargon, unless the user is clearly technical or asks for details, in which case use `checks` (each has a technical message and fix). Blocking AI search or user-request bots hurts AI visibility; blocking training bots is a policy choice, so present it neutrally. If scanBlocked is set, score is null: either the site's bot protection blocked this scanner, or the site's robots.txt doesn't allow it on this page. Say the page couldn't be analyzed, don't present it as a 0 or as proof the site is broken, and report the checks that did run. The scanner identifies itself as AgentReadyChecker, follows robots.txt, and only fetches the page asked for plus robots.txt, llms.txt and the sitemap.

scan_site