The ModularDot scanner applies 16 deterministic checks to one public homepage and two root files. It returns the score, every pass or failure, the evidence it recognized, and the highest-impact gaps. It does not ask an AI model to judge your site.
The score is a description of explicit inputs under a declared method—not a forecast of what an answer engine will do.
The service follows at most three public redirects, reads up to 1.5 MB of returned HTML, and checks whether root /robots.txt and /llms.txt URLs respond. It does not log into sites or execute their client-side interface.
Scan URLs and results are not intentionally stored in ModularDot's lead database. Short-lived Cloudflare infrastructure and security logs may contain request metadata. Do not submit confidential URLs or information.
Pattern checks can miss equivalent language or recognize weak content. A pass means a signal was detected—not that it is accurate, persuasive, legally sufficient, or likely to earn a citation. The scanner does not measure model answers, rankings, traffic, leads, conversion, or revenue. Read the privacy statement and complete methodology.
CitedIndex publishes a dated, independently researched listing of the scanner, including its limitations and source notes. ModularDot does not control the assessment.
The paid Fix Pack adds human review, exact page recommendations, three answer blocks, supported JSON-LD, and an implementation priority plan.