Besides propagating in target space like a point particle, a closed string is also able to wind around non-contractible circles. A direct consequence thereof is T-duality. In the textbook example, it identifies the closed string dynamics on a large and a small circle by interchanging its winding and momentum modes. Patching a background by such dualities clearly goes beyond the notion of conventional geometry. However, there are extensive efforts to embed them into a Framework called string geometry. It provides access to a vast number of new backgrounds with intriguing phenomenology, like e.g. the possibility to obtain de Sitter vacua. Double Field Theory (DFT) is the most promising approach to describe these backgrounds and their properties. But still, it is closely related to the torus. I will present a theory based on Closed String Field theory starting from a Wess-Zumino-Witten model which goes beyond the torus. It plays an important role in clarifying the recent confusion about different constraints in DFT. Furthermore it allows to uplift a large class of new backgrounds to string theory. These backgrounds are not T-dual to any geometric ones.