In a sense, probabilistic programming languages are automatic (approximate) integrators. Programs can represent arbitrary measures, and the back-ends provide various methods of estimating integrals with respect to them.