Vizzly Visual Testing

Hey Product Hunt 👋 I'm Robert, and I built Vizzly. Four years at Percy taught me a lot about visual testing — including everything I wanted to do differently. When I came back to the space two years later and saw nothing had changed, I started building. The tools out there give you a "different" or "same" answer, maybe a diff percentage, and call it a day. That's fine for catching obvious regressions, but it doesn't help you build visual quality into how you actually work. It's always this separate thing you do after development, not during it. So I started from scratch. Built a Rust diffing engine called Honeydiff that's 12-16x faster than the alternatives, with spatial clustering that tells you where changes happened and perceptual color science that matches how your eyes actually see. Then I built a local TDD workflow around it — run vizzly tdd, get instant visual feedback while you code. No sign-up. No cloud. Just you and your UI. That local-first approach was non-negotiable. I wanted developers to catch visual issues while they're still in the flow of building — not three days later when CI finally tells them something broke. When you need team collaboration, the cloud is there. Position-based comments, @mentions, review workflows, the whole thing. But it's optional, not a gate. We've been shipping constantly since June. The changelog has its own dedicated page now, hot spot regions that learn where your dynamic content lives, real-time SSE so the dashboard updates instantly, a Swift SDK for iOS teams, MCP integration so AI assistants can actually debug your visual test failures. It's been a lot. I'd love to hear what you think. What's working? What's missing? What would make visual testing something you actually want to use? Back to building. 🐻 😃

Congrats on the launch!