Before releases I spent hours, sometimes days, babysitting flaky tests. Our team ended up spending ~40–50% of its time maintaining test scripts and fixing issues that could have been caught earlier. Existing tools (Appium, etc.) rely on brittle selectors and hard-coded scripts that break as soon as the UI changes or any random popup appears during testing.
So I built a different approach from the ground up: test like a human. Describe the flow you want to test in plain English, and a Vision+Intent agent executes that flow on real Android and iOS devices. The agent understands context (if a popup appears it dismisses it and continues), maps intent to UI actions, and aims to make test definitions resilient across platforms.
It’s early and opinionated. We optimized for:
1. drastically reduced test maintenance 2. fewer false failures in CI 3. a single test definition that works on both Android and iOS
We’re actively looking for feedback from people who’ve fought mobile testing: what’s still broken for you, and what would make this genuinely useful? If you’ve got flaky suites, device-farm headaches, or cross-platform parity pain — I’d love to hear specifics.
