My goal with this report was to investigate the viability of creating a personalized spell checker using a neural network. The idea is that the network would be able to train using a user’s personal spelling mistakes and a spell checker with high probability of correcting mistakes for that particular user would be created. After some research I found that the best neural network architecture to use when working with text based problems like mine was a sequence-to-sequence architecture. So I built a fairly simple sequence-to-sequence model, trained it on a small dictionary, and then ran a number of tests on it for spelling errors I thought to be important for determining its value as a personalized spell checker. Overall, the network preformed fairly well but there were a few areas that needed some improvement. In the end I concluded that the concept of a personalize spell checker is probably viable but some more research would need to be done to try and strengthen the current model.