In this paper we introduce a novel approach to formal description of non-verbal vocal gestures and implementation of a general recognition system according to the formal description. It was successfully applied to existing non-verbal vocal applications such as mouse emulation by humming or Tetris. Non-verbal vocal interaction (NVVI) is a proven assistive technology and our system accelerates the process of building the NVVI applications. The system is targeted to both designers and users with disabilities who use the recognition system incorporated in an assistive application. Among others, the formal description is capable of providing continuous control of user interface elements and solving inaccuracy in the pronunciation of vocal gestures.