About Vizzly
I built Vizzly to give humans and coding agents the same visual regression testing so design decisions stay grounded in the actual product.
I’m a software engineer who likes building accessible interfaces, reliable product flows, and the tooling that helps teams trust what they ship.
After years working closely with customers at Percy, I learned that screenshot tools only become useful when they preserve enough review history for real decisions. Vizzly is my answer to that problem.
The site and product now lean on visual regression testing and BearDen-inspired local tokens, so the language stays grounded in the workflow instead of an older design-system label.
Outside the keyboard, I’m usually wrenching on my 1995 NA8 Spec Miata, tinkering with hardware, or trying out a new tool idea that eventually wants to become a product.
in UI review
customer conversations
conferences and community talks
Why I’m building Vizzly
Agents need approved baselines
Coding agents can move quickly, but they still need approved baselines and review history before they can make good UI decisions.
Coverage should be encouraged
Broad visual coverage is good behavior. Vizzly keeps the review history rich enough that screenshot volume feels like part of the workflow, not the headline.
Collaboration is the point
Visual review works best when designers, PMs, and developers are all looking at the same evidence.
Ready to try Vizzly?
See what visual regression testing feels like when the product, the docs, and the visual language all point at the same thing.