Acoustic censusing methods offer an efficient and noninvasive method to monitor populations of dense aggregations of animals on a large scale. Large, dense colonies of bats are difficult to census, and current monitoring methods are often invasive and time consuming. Using acoustic recordings collected during emergence, we developed a linear regression model that uses multiple measures of acoustic energy to estimate the number of bats per second. The model was trained using data from audio recordings collected at seven cave locations and two species of bats, Brazilian free-tailed bats (Tadarida brasiliensis) and gray bats (Myotis grisenscens). Visual counts from synchronized thermal videos were used to ground truth the model. Environmental factors, such as distance from microphone to emergence, were integrated to create flexible model that is widely applicable across species and habitats. This method can be extended to other species that vocalize in large aggregations, such as sea birds and frogs. Ultimately, this model can be integrated into autonomous monitoring stations to monitor species in real time. It is imperative that efficient and noninvasive methods for monitoring bat populations are developed to increase population data available for informing crucial management decisions.