AI search readiness is a practical review of whether a page exposes usable content, clear discovery signals and the crawler policy you intend. Screpy retrieves a public page and associated files, then shows the evidence behind each check. It does not calculate a speculative visibility score or claim to predict citations.
Start with the page you want discovered
Enter a full public HTTP or HTTPS URL. A domain alone checks its root page; use an article or product URL when that is the content you care about. The report keeps the final page address so you can distinguish a redirect from the original input. It does not crawl every page on the domain.
Content checks use source HTML without running scripts. They inspect the head title, H1 count, declared canonical, noindex directives, main-text volume, internal links and JSON-LD syntax. An SVG title is not counted as a page title. If essential text appears only after scripts run, investigate it with On-Page SEO Checker and a rendered-page review before assuming the content is absent for every crawler.
Read search policies separately from training choices
The robots table evaluates Googlebot and nine named AI bots or content-use controls against the tested address. It separates search, training, user-requested visits and content-use policy. An intentional training restriction is not automatically an SEO error. The OpenAI bot documentation illustrates why search and training policies must be considered separately.
A matching rule describes published policy, not real network access. This report does not impersonate every bot or inspect your traffic logs. User-triggered agents can have different documented robots behavior; read the behavior column before treating an allowed or blocked row as a prediction. For a focused policy inspection, use AI Bot Access Tester.
Treat discovery formats as evidence, not requirements
The checker requests robots.txt, llms.txt and a Markdown response, and inspects the first same-origin sitemap declared in readable robots.txt. It does not silently try another sitemap when none is declared. Optional formats that are not observed are distinguished from sources that could not be checked.
Google's AI features guidance does not require a special AI text file or special schema to appear in its AI search features. Missing llms.txt or Markdown is therefore not a critical failure. If you choose to publish llms.txt, use llms.txt Validator for a dedicated structure and link review. A parseable JSON-LD block also does not establish eligibility for a rich result; investigate syntax with Structured Data Checker.
Act on complete observations and keep unknowns visible
Each check identifies its source, observed value, status and a practical next step. Errors and warnings appear first. The report uses no 0–100 readiness score: access, source content and optional discovery formats answer different questions and should not be collapsed into an invented probability.
The analysis is limited to five sources, fifteen HTTP steps and 8 MiB overall, with six seconds per retrieval and a thirty-second API deadline. Individual parsers also apply their own size and item limits. A timeout or incomplete source remains unknown; it is not a clean result. Download the JSON report to retain the check time, source responses and coverage with your findings.