The Front-End is implemented by HTML, CSS and javaScript. Django is in charge of handling the requests and communicating with the database. Machine learning techniques are used to conduct the speaker verification.

There are three main components of the algorithms, feature extraction, speaker modeling and feature matching. Feature extraction is used both in entering and matching processes, where entering is to train a voice model using as much audio as possible from one single person, whose features are identified and extracted by some machine learning algorithms like MFCC.

Those extracted features are represented in high-dimensional vectors, which would be simplified in terms of dimension and modeled using statistical model such as HMM. The result is stored in database.​ Feature matching process requires specific algorithms as well, which produces similarity between the uploaded audio and the voice model in the database.​ Decision of accepted or rejected is made according to the similarity number generated.​ In the database, user details and paths to models files are stored.