Multi-ecological integration of visual, auditory, text, and other modal information can break the input and output limitations of a single-mode, and thus get closer to the real human habits. For AI interactive products, only the comprehensive utilization of multiple ecological information can understand human needs more accurately.