ARTNet is a deep learning-based model to predict potential bacterial ADP-ribosyltransferase toxins. For each query sequence, we apply three ARTNet models trained on pos_art_346, pos_art_346_random and pos_whole to produce an ensemble prediction result. We offer three modes: comprehensive, medium, and strict to report positive sequences supported by at least one model, at least two models, and all three models, respectively. The mode of comprehensive is the default output. Only positive results will be reported. For privacy protection, submitted sequences and predicted results will be deleted from the server 3 days after analysis.