minitap Information

Mobile teams shipping with Minitest don't write or maintain test code. Our agent owns the test suite end-to-end: it reads your application, runs the user's job the way a real user would, and tells you in Slack whether the release is safe to ship. What scripts cannot see, the agent does. While a run is in progress, it inspects application logs in real time and watches CPU and memory usage the way an engineer reading a profiler would. That extends what your tests catch beyond functional regressions, into memory leaks, performance regressions, and the non-deterministic behaviour of AI features in the model-driven parts of your product. UI and UX issue detection is shipping next. Minitest is built on mobile-use, our open-source agent SDK that leads Google DeepMind's AndroidWorld benchmark. Mobile is where the cost asymmetry is sharpest, the in-house QA function is usually missing, and the App Store penalty makes the math break first. See the agent in action at minitap.ai/demo. Book a call from there when you want it on one of your apps! Founded in 2025 by Nicolas Dehandschoewercker and Luc Mahoux-Nakamura. Team of 10 engineers based between San Francisco and Paris (Station F). Backed by Moxxie VC, Mercuri VC, EWOR, and founders from Hugging Face, SumUp, Adjust, FlixBus, Last.fm, and World.