About
Evidence before advice
Lightmark is an independent website diagnostic operated by Adam. It measures what a public fetch can prove, then turns each failed check into a concrete implementation brief.
What Lightmark measures
The current scanner runs 39 checks grouped into 11 scorecards: AI readiness, technical SEO, performance, agent readiness, accessibility, security, privacy, reliability, trust, content quality, and conversion UX. Each scorecard remains visible so a strong result in one system cannot hide a blocker in another.
Every scan compares 2 public fetch profiles, a normal browser user agent and GPTBot. Lightmark inspects response status, raw HTML, crawler parity, discovery files, metadata, structured data, accessible names, scripts, headers, trust links, content structure, and conversion paths.
From diagnosis to cure
A finding is useful only when it can be acted on and verified. Lightmark records the observed evidence, recommends the smallest relevant change, defines a done-when condition, and generates 1 prioritized Improve prompt that a developer or coding agent can apply to the real repository.
What the score does not claim
A single public scan cannot prove field Core Web Vitals, historical uptime, private analytics, conversion rates, authenticated behavior, or whether a model cited the site. Lightmark reports those as limitations instead of inferring data it did not observe. Its accessibility, security, and privacy checks are review evidence, not certifications.
Operator and source
Adam operates Lightmark and maintains the public source repository. Product questions and implementation requests can be sent through the contact page; code-level issues can be reviewed in the GitHub repository.