Simulating phenomena that take place in different geometric scales often requires substituting high-fidelity models with lower-dimension simplifications. Common examples include 1D-3D flow simulations in hemodynamics, 1D-3D nuclear reactor thermohydraulics, 2D-3D free surface flows, and more. While the range of such problems is wide, solutions and software are often application- and code-specific. preCICE provides an application- and code-agnostic solution for coupling existing models for multi-physics simulations. Apart from a core library for coupling arbitrary codes written in C++, C, Fortran, Python, Matlab, or Julia, preCICE also includes further components that make up a sustainable coupling ecosystem. Ready-to-use, add-on components such as the OpenFOAM-preCICE adapter and further integrations to established codes allow users with limited coding experience set up efficient and accurate simulations for fluid-structure interaction, conjugate heat transfer, fluid-fluid coupling, and more. With recent additions to the core library and to the OpenFOAM adapter, users are now able to compose mixed-dimension simulations out of 1D and 3D fluid models, without one model being aware of or specific to the dimensionality of the other. This talk will present an experimental version of the geometric multiscale feature of preCICE, and related developments and features in the OpenFOAM adapter that make this solution possible, flexible, sustainable. These developments enable new use cases for preCICE, such as in the preCICE-ATHLET project, coupling the 1D thermohydraulics code ATHLET to the 3D CFD code OpenFOAM.