Hy-Nomas is an open-source interactive video world model that generates continuous video segments guided by free-form text prompts and camera trajectories. It maintains consistency across long sequences using world memory and runs efficiently on consumer hardware.