AbstractThe onset of sliding motion is conditional on the propagation of rupture fronts that detach the contacting asperities forming a frictional interface. These ruptures, when propagating over a fault surface, are the most common mechanism for an earthquake. Experimentally, the transition from static to sliding friction takes place when a rupture traverses the entire interface. But ruptures can also arrest before reaching the end of the interface. The determination of the mechanisms responsible for rupture arrest is of particular interest for understanding an earthquake's magnitude selection. Propagating ruptures have been shown to be true shear cracks, driven by singular fields at their tip, and fracture mechanics have been successfully used to describe rupture arrest along homogeneous frictional interfaces. Performing high temporal resolution measurements of the real contact area and strain fields, we demonstrate that the same framework provides an excellent quantitative description of rupture arrest along interfaces with heterogeneous fracture properties and complex stress distributions at a macroscopic scale. This work unravels the different mechanisms responsible for rupture arrest along model laboratory faults. This fracture‐based paradigm opens a window to a wide range of possible consequences for frictional behavior along any two contacting bodies; from the centimeter scale to the scale of natural faults.