In this project, a model able to do optical character recognition with a deep learning approach is implemented. The starting point is the analysis and exploration on the current state-of-the-art in computer vision. Following the structure of a Neural Network is proposed, analysed and implemented. Finally this model is tested with several use cases.