MapPred is a web server for deep learning-based prediciton of protein inter-residue contacts and distances. For each query sequence, a multiple sequence alignment (MSA) was constructed. Deep residual neural networks are used as the engine for training and prediction, with covariance features derived from the MSA. MapPred is shown to be comparable or superior to the state-of-the-art methods. The output inlcudes the predicted contact map, distance maps and the distance distribtuion.