With the democratization of EEG devices, the possibility of on-the-fly EEG analysis has grown, although some of the major locks are the prices of this equipment and the reliability of signal classification. In this study, we propose to use cost-effective materials to develop reliable hardware and to train machine learning algorithm to prove that it is possible to classify EEG signals on open environments. It was decided early in the process to classify EEG signals corresponding to actions as it can be used to help people with medical difficulties better interact with their environment. In addition, long-term monitoring of the subject's actions through their EEG signals could be useful to keep an accurate record of their movements.