Over the last 8 months, I’ve been learning the SaaS stack and building small side projects. One thing I consistently struggled with was conversion rate optimization. Most guidance felt subjective, and A/B testing wasn’t realistic with low traffic.
I built Crovise to explore whether static analysis could be used to generate useful CRO hypotheses. Instead of tracking users or predicting outcomes, it analyzes the HTML and DOM structure of landing pages: semantic tags, hierarchy depth, CTA placement, and common structural patterns.
The idea is closer to static code analysis than analytics — surface potentially weak or interesting structures that are worth testing.
Technically, the system is rule-based by design and built with Next.js. The hardest part was translating qualitative UX heuristics into deterministic rules without making them too rigid.
This is an MVP and is currently in a waitlist phase. It works best on simple marketing pages and struggles with complex SPAs or highly dynamic content. False positives are expected.
Thanks for reading.