Tissue bending is a key part of the animal morphogenetic toolkit that can happen at different scales, from small invaginations to larger embryo-level deformations. An example of the latter is the ventral folding of the tail that occurs in numerous animal embryos. However, the mechanisms underlying tail bending are still incompletely understood; one idea is that bending is passive, a result of confinement of the embryo within the chorion. Now, Bo Dong and colleagues use the simple chordate Ciona as a model system to investigate these problems. They first show that bending is in fact an active process: it still occurs in dechorionated embryos, and is autonomous to the tail. Active actomyosin is enriched ventrally in the notochord and, as tested both pharmacologically and genetically, required for bending. In the overlying epidermis, proliferation is higher in dorsal compared to ventral sides, and inhibition of proliferation leads to a failure in bending. Atomic force microscope measurements reveal that notochord is the stiffest tissue in the tail, and the authors then built a mathematical model to test the relative contributions of each tissue: this modelling suggests that ventral constriction of the notochord is sufficient to drive bending. Thus, tail bending is an autonomous morphogenetic process driven by asymmetric cell behaviours.