Developer tools at scale. Agents next.

I spent years building developer tools for humans. Runme started as a simple idea — interactive runbooks that actually work — and grew into a CNCF Sandbox project used across hundreds of organizations. The goal was always the same: make complex technical workflows feel natural.

Agents are the next UX frontier. But here's the problem: everyone's building the models, the frameworks, the infra. No one's testing the experience. We optimized browser UX for decades — automated testing, observability, performance monitoring. For agents? We're flying blind.

We're building visr because agent UX needs the same rigor we've applied to browser-based webapps. If agents are going to handle real workflows, we need to know when they break, where they confuse users, and how to make them better.