A test/run program receives as input a list of identifiers for source pages referencing applets to be tested or run. The test/run program creates an array of the identifiers, together with parameters for each identifier, web browser to run the test under, and a number of times the source page is to be reloaded and the applets re-run. For each source page, and for each reload of a given source page, the test/run program starts the specified web browser process, loads the designated source page, and starts a fresh runtime environment for the applet. Support for a test class within the test/run program allows the applets to write success, failure, or informational results to an output file and to exit the web browser process when complete. Where a native implementation of the test class is employed, special security permissions need not be specified and the test/run program need not necessarily be run locally. In exiting the web browser process, the applets write a marker file to indicate that the applet run is complete, which the test/run program detects. Multiple applets may be automatically and repetitively loaded, each with a fresh runtime environment in a new web browser application, for testing of the applets or repeat execution of the applets changing system properties.