The stability of a webchat is dependent on numerous factors, ranging from quality of code and browser compatibility, to the stability of the infrastructure and the way it handles network instabilities. To make a sensible comparison between webchats or to measure changes in stability with a new version of a webchat, there needs to be some kind of standardized testing. In this talk I will show how I have set up an automated testing environment that automatically simulates numerous potential problems, including network instability.