SuperNNova performs photometric classification by leveraging recent advances in deep neural networks. It can train either a recurrent neural network or random forest to classify light-curves using only photometric information. It also allows additional information, such as host-galaxy redshift, to be incorporated to improve performance.