Despite offering deeper insights into models and more trustworthy predictions, Uncertainty Quantification (UQ) remains much less widespread than purely deterministic simulation. In particular, the most advanced UQ methods are not routinely applied to computationally challenging numerical models, even though that's where they offer the greatest benefits. We believe this is largely due to technical complexity and lack of separation of concerns when combining state-of-the-art UQ, models and high performance computing (HPC) capabilities. Addressing these issues, we present UM-Bridge, a universal software interface for coupling any simulation code to virtually any UQ software. Easy to use integrations for C++, Python, R, MATLAB and Julia as well as several UQ packages like PyMC, MUQ, QMCPy and Sparse Grids Matlab Kit are available. Inspired by microservice architectures, UM-Bridge enables containerization of models. Making use of that, we are building the, to our knowledge, first library of ready-to-run UQ benchmark problems in a community effort. Further, UM-Bridge provides platforms for scaling up models on cloud or HPC systems, enabling even prototype-grade UQ applications to offload costly model evaluations to powerful compute clusters. Finally, we demonstrate how UM-Bridge enables rapid development of advanced UQ algorithms, straightforward parallelization and application to challenging real-world problems.