The application of deep-learning-based seismic phase pickers for earthquake monitoring has surged in recent years. However, the efficacy of these models when applied to monitoring volcano seismicity has yet to be evaluated. Here, we first compile a dataset of seismic waveforms from various volcanoes globally. We then show that the performances of two widely used deep-learning pickers deteriorate systematically as the earthquakes’ frequency content decreases. Therefore, the performances are especially poor for long-period earthquakes often associated with fluid/magma movement. Subsequently, we train new models which perform significantly better, including when tested on volcanic earthquake waveforms from northern California where no training data are used and tectonic low-frequency earthquakes along the Nankai Trough. Our model/workflow can be applied to improve monitoring of volcano seismicity globally while our compiled dataset can be used to benchmark future methods for characterizing volcano seismicity, especially long-period earthquakes which are difficult to monitor.