An automated off-body Mach cone aligned structured curvilinear grid generation procedure is presented for near-field computational fluid dynamics simulations. This procedure combines output-based indicators and mesh redistribution to perform anisotropic mesh adaptation while maintaining Mach cone alignment. Automation is achieved through a novel direction-based adaptation indicator formulation. The adaptation procedure is demonstrated on the JAXA Wing Body geometry and X-59 C608 demonstrator models from the Second and Third AIAA Sonic Boom Prediction Workshops, respectively. It is demonstrated that anisotropic mesh adaptation may result in a greater than 50% reduction in the resource usage required to achieve the same level of accuracy as uniform and user-constructed Mach cone aligned grids for near-field pressure signatures, ground-level overpressure signatures, and loudness metrics.