视频 图片
登录

Monkeys Rob Her Twice

观看2,037
点赞104
评论70
分享8
收藏17
引用0
Create ultra-realistic 15-second vertical iPhone footage during a sunny summer afternoon at a quiet Japanese mountain village rest stop. Use the provided reference image ONLY for the main woman's identity. Keep her facial features, hairstyle, skin tone, outfit, and overall appearance consistent throughout.

The footage should look like a friend casually recording a real travel moment. Natural handheld smartphone movement, imperfect framing, realistic autofocus, authentic daylight, and spontaneous reactions. The scene must feel unscripted and believable, not cinematic.

0–3 seconds:
The woman stands beside a Japanese vending machine, opening a cold plastic drink. A Japanese macaque monkey sits on a wooden railing nearby, watching her. She notices the monkey briefly, then looks back at her drink.

3–6 seconds:
Without warning, the monkey quickly jumps down, snatches the bottle from her hand, and darts onto the nearby railing. The woman instinctively reaches for the bottle, then freezes with a genuinely surprised expression. The camera quickly turns to follow the monkey.

6–9 seconds:
The monkey sits on the railing and investigates the bottle with both hands, trying to open it. The woman takes two cautious steps closer, watching with a confused, disbelieving look. She remains composed, with only a slight natural smile.

9–12 seconds:
The monkey drops the bottle cap onto the ground. The woman bends slightly to look at it, then glances back at the monkey. Suddenly, a second macaque approaches from behind and reaches into her open tote bag, pulling out a wrapped snack.

12–15 seconds:
The woman immediately turns at the sound of her bag rustling and discovers the second monkey escaping with her snack. She quickly pulls her bag toward herself, looks from the monkey to her now-empty hand, and gives the camera a silent, incredulous look. The person filming quickly pans toward the escaping mon

✨ 生成 原始帖子 在 Hitboard 打开

如何使用这个提示词

复制提示词粘贴到 MiniMax。分镜结构、镜头运动与时间点请按原文保留,这是产出该结果的关键;只改主体、场景或风格用词让它成为你的,并附上提示词提到的参考图。第一次生成若跑偏,优先缩短提示词,而不是再加指令。