The typical real-time microseismic monitoring workflow involves using seismic sources at known locations, such as perforation shots, string shots, or ball-seating events, to calibrate the velocity model. In some cases, none of these sources are available, they are of poor quality, or their coverage of the rock volume of interest is limited (e.g., in horizontal monitoring configurations). Crosswell seismic data provide a great opportunity to calibrate a velocity model for use in real-time microseismic monitoring when optimal calibration sources are not available. We tested two data sets with different well configurations and available calibration data to demonstrate the use of crosswell sources for velocity-model calibration. The first data set, which includes both crosswell sources and perforation shots, shows that using crosswell source shots for velocity-model calibration results in event locations as good as locations based on perforations. In a second example, we establish the feasibility of using crosswell sources to extend the depth coverage of velocity-model calibration for a horizontal monitoring configuration. In both cases, using crosswell sources can provide a means to reduce uncertainty in the velocity model and improve subsequent event locations.