Ask anything and stream answers from the distributed inference mesh.