Its a stm32f4 discovery based project. It recognizes 6 spoken words and selects appropriate output - View it on GitHub