At the continent-scale, models of Moho depth based on seismic estimates alone can be inadequate due to irregular or sparse data. Gravity-based Moho modelling provides better coverage, however, the methods used are typically hampered by an inability to explicitly honour seismic constraints and are also limited by over simplistic model conditions, e.g. laterally-homogenous layering. I present a new method to generate a continent-scale Moho model, based on the constrained inversion of free-air gravity data. This method explicitly honours seismic Moho estimates and accounts for a laterally heterogeneous crust and mantle. Resolution and sensitivity testing shows that, for wavelengths greater than 200 km, crustal density and Moho depth are recovered with reasonable accuracy, +/- 30 kg m(-3) and +/- 3 km respectively. MoGGIE uses a six layer model incorporating ocean, sedimentary basin, upper crust, lower/oceanic crust, eclogitised crust and mantle. Inversion variables were the density of the crustal layers, constrained by a standard density model, and the depths to intra-crustal boundaries and the Moho, constrained by 230 seismic depth estimates. The results demonstrate that a balanced approach to seismically-constrained gravity inversion has the capability to generate detailed and well-constrained models of the Moho and crustal density at the continent-scale.