The paper is discussing the phoneme level interpretation of sound signal to form the words in Malayalam language. Natural language processing with advanced algorithms adopted here to ensure the accuracy and optimization of the processing. Analog to digital conversion of voice signal is taking numerous methodologies for a perfect processing. Training a new system with these techniques to test whether the given commands belongs to dataset or not.

In this paper, a new method to read and recognize the voice command given in Malayalam language with the help of advanced language processing algorithms. Tensor flow is taken as the tool for voice processing in python language and various methods to train the signals and test them in second level. The acoustic model will check whether the signal can be recognized by these methodologies for further actions to be taken down. The accuracy is also to be considered with large data set.