Phase separation in equilibrium systems, think to a fluid separating in macroscopic vapour and liquid regions, is described by the \phi^4 field theory (Model B in the Hohenberg-Halperin classification). As it is very well known, this theory is built upon a gradient expansion that includes all the terms not forbidden by microscopic symmetries. Phase separation happens also in many non-equilibrium contexts, where time-reversal symmetry is violated. This happens in systems as diverse as active systems, the interior of cells or shaken granular matter. I will present our recent efforts building the first consistent field theory for phase separation when time-reversal symmetry is broken locally. Surprisingly, the ensuing phenomenology is deeply affected both quantitatively (universality class) and qualitatively. Part of our discussion will be focused on active systems, those where each degree of freedom is able to transform non-thermal energy into motion (such as animals or bacteria), but our field-theoretical results are expected to apply much more broadly.