A method for aligning coordinate systems from separate augmented reality (AR) devices is described. In one aspect, a first device accesses first pose data from a first Visual Inertial Odometry (VIO) system of the first device. A camera of the first device captures an image of a machine-readable code that is displayed on a display of a second device. The second device encodes the machine-readable code with second pose data from a second VIO system of the second device. The first device decodes the second pose data from the machine-readable code, and determines a relative pose between the first device and the second device based on the first pose data and the second pose data.