Incorporating climate model information in rainfall frequency and probable maximum precipitation (PMP) analysis has been of great interest given climate change influences on extreme storms. Recent efforts have used dynamic weather models to estimate PMPs, but were limited to small domains and specific storm types. We propose a semi-dynamic stochastic rainfall generator that can simulate extreme rainstorms at continental-scales at 6-hour and 0.03° resolution for more than 10,000 synthetic years, conditioned on atmospheric variables from global climate models. The generator focuses on winter and spring rainstorms in the Mississippi watershed caused by strong water vapor transport from the Gulf of Mexico—the types of storms that produce large floods in the basin. The generator simulates spatiotemporal noise fields that replicate observed rainfall structures, which are then converted to rainfall fields through parametric rainfall distributions conditioned on large-scale atmospheric variables from the Community Earth System Model 2. For a given set of atmospheric fields, the generator can produce multiple realizations of extreme rainfall scenarios, reflecting the inherent uncertainty in local rainfall dynamics associated with a given large-scale atmospheric environment. Validation shows that the generator produces realistic rainstorms that are visually and statistically consistent with observations. Our proposed approach incorporates climate model information via conditional simulations, improving the representation of rainstorm arrivals, space-time structures, and statistical properties. The generator provides a wide range of extreme precipitation scenarios that can be used both for rainfall frequency and PMP/PMF studies in the Mississippi River Basin and can be easily extended to other rainstorm types and locations. Compared with conventional extreme value analysis, our results show a bounded upper tail of extreme rainfall, consistent with the notion of PMP. This study also suggests that for continental-scale river basins, it is a critical sequence of extreme rainstorms—rather than a single event—that constitutes the PMP/PMF in the basin.