Abstract: This paper presents a brief review of speaker & language recognition system using Hidden Markov Model (HMM).For accurate personal identification systems the use of biometric is preferred over the security system implemented by password and pin number. Speech recognition was biometric feature of voice of the speaker. Different speaker have different voice characteristics, these different characteristics are achieved by extracting feature vectors as MFCC from speech. The brief history of the hidden Markov Model explain about voice signal and the evolution of the HMM is done in Google Web API & this is implemented using low cost raspberry Pi. With the help of these techniques the performance has increased. The outputs of system are through speaker.

Keywords: Speech recognition (SR), MFCC, Hidden markov model (HMM), and Raspberry pi, Google Web API, Microsoft Web API.