Advances in simulation methodologies, code efficiency, and computing power have enabled larger, longer, and more-complicated biological membrane simulations. The resulting membranes can be highly complex and have curved geometries that greatly deviate from a simple planar state. Studying these membranes requires appropriate characterization of geometric and topological properties of the membrane surface before any local lipid properties, such as areas and curvatures, can be computed. We present MemSurfer, an efficient and versatile tool to robustly compute membrane surfaces for a wide variety of large-scale molecular simulations. MemSurfer works independent of the type of simulation and directly on 3D point coordinates. As a result, MemSurfer can handle a variety of membranes. Using Delaunay triangulations and surface parameterizations, MemSurfer not only computes common lipid properties of interest but also provides direct access to the membrane surface itself, allowing the user to potentially conceive and compute a variety of nonstandard properties. The software provides a simple-to-use Python API and is released open-source under a GPL3 license.