1272 hand-tested AI prompts across image, video, and music — each with full context, variables, and a sample output. Content is authored in Traditional Chinese; the UI here is English.
把人像換成水彩／3D／動畫等風格，臉部一致不走鐘。