Hi HN, I’m Adam. I built Crovise, a tool that performs static analysis on landing pages to generate CRO hypotheses.

I started this after building a few small SaaS projects where improving conversion rates often felt vague and subjective. Most advice was generic (“make the CTA clearer”) or required long A/B tests that weren’t realistic with low traffic. I wanted something more deterministic and engineering-driven.

Crovise analyzes the HTML structure and DOM hierarchy of a landing page to surface potential conversion issues and testable hypotheses. It looks at semantic structure, element placement, hierarchy depth, and common patterns seen in high-converting pages — not just visuals or copy tone.

Technical notes:

- The analysis engine is currently rule-based, not ML-heavy - Built with Next.js - Translating qualitative UX heuristics into deterministic rules was the hardest challenge

This is an MVP in a *waitlist / early-access phase*. It works best on simple marketing pages; complex SPAs or dynamic content are a known limitation, and false positives are expected.

I’m 16 and have been learning the SaaS stack for 8 months. I’d really appreciate feedback on:

- Whether static analysis is a reasonable approach for CRO - Which signals feel useful vs misleading

Waitlist/demo: https://crovise.netlify.app

Thanks for taking a look.