This set of APIs aimed on processing of the speech files. Main feature is a noise suppressor, capable of removing virtually any type of noise from speech streams. It handles passing cars, restaurant background noise, crying children or whistling birds. Suppressor automatically adjusts gain (volume) of the signal.
Classification API (prototype) allows to segment files on speech regions, annotating each region with set of attributes. Currently it has language tag and we will add more soon, like speaker dependent stable hash, age, sex and sentiment. Stay tuned.