This paper presents the implementation of a versatile MATLAB application focused on acquiring, visualizing and storing the electromyographic (EMG) signals read by the sensors of the Myo Armband device, where it is possible to perform signal processing by means of a predetermined function by the user, in order to be able to build databases of both raw and processed EMG signals. It also includes an option to perform real-time tests of convolutional neural networks that have been trained with the acquired databases. To test the application, it is presented a basic example of acquisition and processing of the acquired signals for the recognition of 2 hand gestures, using Power Spectral Density as feature extraction function, and with the feature maps obtained through the extraction, the training of a convolutional neural network is performed, getting a 95% accuracy of recognition, and additionally, the validation is performed through real-time tests within the application, demonstrating the usefulness and performance of the developed application.