This research proposes a prototype to identify speech and language for multiple languages. Acoustic features such as MFCC and delta-MFCCs are used for this work. At the back-end i.e. at the classifier stage Artificial Neural Network (ANN) and radial basis function are used. ANN is based on supervised learning approach for the purpose of speech and language recognition. Back propagation algorithm is used for training ANN. ANN model attempts to classify the input speech signals on the basis of a set of words and languages. This research is based on four Indian languages Hindi, Telugu, Sanskrit and English are used.