We updated our architecture yesterday with some major improvements! After many iterations and various approaches to reducing the processing overhead while maintaining the same accuracy, we've come to a solution that we are extremely excited about. Our new detection methods are approximately 10x faster than the old system which means we now have so much room for acitivites. We'll immediately be putting energy towards crawling a larger part of the web and releasing a chrome browser extension.
Our old system hasn't been completely replaced and still serves as our back-up detection algorithm should the new system fail to make a confident detection. Currently, approximately 89% of the detection requests are being fulfilled with our new system, almost 11% fulfilled by our old system and less than 1% finishes without a detection (who out there is trying to find the hosting provider of 127.0.0.1?).