The theory of group actions on CAT(0) cube complexes has exerted a strong influence on geometric group theory and low-dimensional topology in the last two decades. Indeed, knowing that a group G acts properly and cocompactly on a CAT(0) cube complex reveals a lot of its algebraic structure. However, in general, "cubulations'' are non-canonical and the group G can act on cube complexes in many different ways. It is thus natural to try and formulate a good notion of "space of all cubulations of G'', which would prove useful in the study of Out(G) for quite general groups G. I will describe some results in this direction, based on joint works with J. Beyrer and M. Hagen.[-]