The increasing attractiveness of Luxembourg as a place to work and live puts its land use and transport systems under high pressure. Understanding how the country can accommodate residential growth and additional traffic in a sustainable manner is a key and difficult challenge that requires a policy relevant, flexible and responsive modelling framework. We describe the first fully fledged land-use and transport interaction framework (MOEBIUS) applied to the whole of Luxembourg. We stress its multi-scalar nature and detail the articulation of two of its main components: a dynamic demographic microsimulation at the scale of individuals and a micro-spatial scale simulation of residential choice. Conversely to traditional zone-based approaches, the framework keeps full details of households and individuals for residential and travel mode choice, making the model highly consistent with theory. In addition, results and policy constraints are implemented at a very fine resolution (20m) and can thus incorporate local effects (residential externalities, local urban design). Conversely to fully disaggregated approaches, a linkage is organized at an intermediate scale, which allows (1) simplifying the generation and spatial distribution of trips, (2) parallelizing parts of the residential choice simulation, and (3) ensuring a good calibration of the population and real estate market estimates. We show model outputs for different scenarios at the horizon 2030 and compare them along sustainability criteria.