Kobiton launched a Claude MCP plugin that lets developers create, debug, and run mobile tests on real devices using natural language inside Claude or other AI coding tools. The podcast covers a live demo showing how Claude can run Appium scripts on real devices in Kobiton's cloud, automatically detect bugs (like keyboard overlap on small screens), and propose fixes. Topics include AI-generated test quality risks (false positives, tests that miss the point), token cost optimization via local models, natural language element selectors for canvas/PDF apps, and the evolving role of QA engineers as AI accelerates code output 10x.
•31m watch time
1 Impression