Configuration is entirely file-based, in `nabiki.json', which _is_ read by a proper JSON interpreter, so formatting is important. The included example file contains descriptions of the accepted options. http://jshint.com/ is quite useful for checking, if you have any difficulty. You will have to supply your own Ethereum address (mine is there as a placeholder - please don't actually mine to it!) and select a server. CUDA devices are numbered starting at 0, which is the "most capable" card according to Nvidia's criteria - not always accurate, but generally at least close. Specific devices can be disabled if desired, and per-card intensity is, hopefully obviously, supported easily. Have no worries about specifying cards that don't exist: such settings are simply ignored. Once your configuration is complete, simply double-click on the miner and enjoy. No more faffing about with arcane batch file syntax, piping commands and so on. Pool mining only, I'm afraid. There is, finally, a ~2.5% dev fee. The implementation is simple and uses a foolproof (though I tested it anyway!) counter: it merely diverts every 41st share to my address. Dev shares are not indicated in your solution counter, but instead trigger a log message containing a counter so that you can verify the ratio and timing. Happy mining!