万圣节 AI 照片
AI Halloween Costume Generator 给你的一张照片换上万圣节服装：在烛光哥特长廊里的经典吸血鬼，或者在起雾秋日森林里的女巫，另外还有五套服装的提示词可以复制。买或做服装之前先用 AI 试穿一下：衣服换了，你的脸、发型、肤色和身材都保留。
更新于 Vidlark 团队整理可用于 Google Gemini · ChatGPT
情侣吸血鬼装
2 张照片 · Nano Banana 2.1 · 40 积分
挑万圣节服装，通常只能猜它穿在自己身上是什么样。AI 万圣节服装生成器让你先看到：上传一张照片，就能看到自己扮成吸血鬼或女巫，站在适合这身装扮的场景里，还没花一分钱买布料或假牙。
每条提示词都把整套服装从领子写到鞋子，并保留照片里的脸、发型、肤色和身材。卡片上的两条是测试过的；下面另外五个点子，狼人、木乃伊、海盗船长、天使和骷髅，也是同样的写法：服装都遮得严实，只有一个人，画面里没有别人。
只要一张你的照片，服装和场景都写在提示词里。
吸血鬼和女巫是卡片用的提示词；另外五条照同样写法写成文字：复制一条，或带着你的照片直接打开去生成。
Halloween costume photo of the person from the photo dressed as a classic vampire: black high-collared cape with a red lining, black waistcoat, white shirt, pale skin makeup and small fake fangs, standing alone in an empty candle-lit gothic hallway; nobody else appears in the scene. Keep the face, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
Halloween costume photo of the person from the photo dressed as a witch: a modest long-sleeved black dress with a high neckline, a tall pointed black hat, a purple velvet cape and a broomstick in hand, standing in a misty autumn forest with lanterns. Keep the face, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
Halloween costume photo of the person from the photo dressed as a werewolf: a torn plaid flannel shirt buttoned up and dark jeans, faux-fur pieces along the jaw and on the backs of the hands, pointed furry ears, amber eyes and small fake fangs, standing alone on an empty moonlit forest road with a full moon behind; nobody else appears in the scene. Keep the face, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
Halloween costume photo of the person from the photo dressed as an Egyptian mummy: loose linen bandages wrapped over a full-length beige bodysuit from the neck to the ankles, a few bandage ends trailing, the face left uncovered with light dusty makeup, standing alone in an empty torch-lit tomb with hieroglyphs on the walls; nobody else appears in the scene. Keep the face, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
Halloween costume photo of the person from the photo dressed as a pirate captain: a long dark-red coat with brass buttons over a white linen shirt, a black tricorn hat, a wide leather belt, tall boots and a brass spyglass in hand, standing alone on the empty deck of an old wooden ship at sunset; nobody else appears in the scene. Keep the face, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
Halloween costume photo of the person from the photo dressed as an angel: a long-sleeved floor-length white robe, large white feathered wings, a soft golden halo above the head, standing alone in an empty misty chapel with colored light through stained glass; nobody else appears in the scene. Keep the face, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
Halloween costume photo of the person from the photo dressed as a skeleton: a long-sleeved black bodysuit with a white bone print from the neck to the ankles and skull face paint, standing alone in an empty foggy graveyard at night lit by a few lanterns; nobody else appears in the scene. Keep the face shape, hairstyle, skin tone and body shape exactly as in the photo. Photorealistic costume party photo, vertical 3:4. No text, no logos, no watermark.
一览
复制提示词是免费的。在这里生成，一张 1K 的 Nano Banana 2.1 图片用 40 积分，所以注册送的 100 积分能做 2 张。月付套餐里 Starter 每张 32 积分，Pro 20 积分。Unlimited 做 Nano Banana 2.1 图片不用积分。
一个成年人清楚的照片，最好是全身，整套服装才能画出来。照片里穿什么没关系：提示词会换掉。
会，这正是它的用途：提示词给你换上服装。你的脸、发型、肤色和身材都照照片保留。
卡片做的是吸血鬼和女巫。下面还有五条可以复制的服装提示词：狼人、木乃伊、海盗船长、天使和骷髅。
能。拿任意一条提示词，把服装的词一件件换成你的；保留那句保持你的脸的话。
能。情侣吸血鬼那张卡片要两张照片（照片 1 和照片 2），给你们俩穿上配套的高领斗篷，在万圣节派对上合影。其他提示词是给一个人的；下面链接的玫瑰万圣节服装页面还有更多情侣装扮。
能。在结果下面点「做成视频」，图片会作为第一帧在图生视频里打开，就能做一段短视频。
一张图大约一到两分钟。会存进「我的作品」，生成时可以离开页面。
看和复制提示词不需要。要在这里生成，就注册一个免费账户：回来时卡片、提示词和你的照片都还在。