Systems and techniques to test a software product in one or more computer systems. In general, in one implementation, the technique includes calling components in a first computer system during execution of a test, recording the calls, and rerunning the recorded calls in a second computer system in order to recreate the behavior of the first computer system during execution of the test.