A Hacker News user posted a demo that pits two sites against each other, built on Cloudflare’s Browser Run and Workers AI platforms. The app showcases a “website vs website” showdown, using the provider’s edge‑runtime to fetch, render, and compare pages in real time for quick visual feedback.

Behind the scenes, Workers AI calls Google’s Gemma 4b model to analyze visual elements. Despite being a 4‑billion‑parameter model, Gemma proves capable of interpreting screenshots and extracting layout cues, enabling the system to judge which site presents a cleaner design and more intuitive user experience today.

The combination of edge execution and on‑device inference sidesteps traditional server latency, letting the comparison happen within milliseconds. Developers can repurpose this pattern for A/B testing visual changes, automated UI audits, or building playful competitive dashboards without provisioning separate back‑ends in the cloud environment today.

Community response on Hacker News has been modest, with ten points and five comments noting the novelty of merging Cloudflare’s serverless stack with a cutting‑edge vision model. The project illustrates how low‑code AI services can power quirky yet useful web tools that run entirely at the edge.