提示词库 · 复制并运行

万圣节 AI 提示词

这里整理了十二条已经可用的提示词:服装、宠物、情侣、幽灵肖像和短视频片段。复制其中一条,粘贴到本页工具中,然后只替换你在意的细节。

12 条配方复制即用图像 + 视频无水印

更新于 2026-10-20 · Crazy Frog AI Video

一张深色桌子,上面放着打印好的提示词卡片、一个水晶球和温暖的烛光

快速解答

一条好的万圣节提示词包含五部分:保留什么、更改什么、主体站在哪里、光线是什么样,以及需要避免什么。多数单句提示词失败,是因为它们只描述了要更改的内容。

本页提示词面向图像编辑模型和图像转视频模型编写。它们会保留面部和人数,再更改服装与场景。

你可以在本页工作室中运行任意提示词。每张静态图片消耗 1 个积分;12 秒视频在 480p 下消耗 15 个积分。

Photos are moderated and only used for this render.

Keep the identity and “avoid” lines. They are what stop the model from inventing a new face.

Frame

Sign in to upload and render. Your first pack starts at $9.90.

效果预览

一位身穿星空长裙、头戴月亮王冠的天界公主肖像
宇宙公主——提示词在上方提示词库中。
带有得体万圣节妆容的维多利亚风格幽灵肖像
幽灵肖像——靠妆容,不靠血腥。
两人分别穿着协调的牛仔与天使服装
牛仔与天使——两张照片、两个身份,不换脸。

五段式提示词骨架

本页每条配方都遵循同一顺序。先写保留内容,因为身份是用户第一眼就注意到的部分。然后是更改、场景、光线,最后是负面清单。如果只记住一件事,就记住这个顺序:模型会更重视前面的分句。

提示词不是愿望清单。每个句子要么保护某样东西,要么改变某样东西。“让它更恐怖”没有给模型可衡量的更改;“把主体放在冷蓝色门廊灯下,身后是雕刻好的南瓜”就可以。

  • 保留:面部、肤色、年龄、身体比例、人数。
  • 更改:发型、服装、妆容、配饰、道具。
  • 场景:房间、街道、森林、工作室、背景物体。
  • 光线:一天中的时间、方向、色温、阴影质感。
  • 避免:多余肢体、融合的面部、文字、标志、水印、血腥。

图像提示词与视频提示词是不同任务

图像提示词可以要求一个全新场景,因为它只需生成一帧。视频提示词必须让这一帧在数秒内保持连贯,所以应该要求更少:转头、眨眼、飘动的雾、缓慢推近。大幅动作——奔跑、跳舞、换装——正是人体结构容易崩坏的地方。

提示词库中最后两条展示了这个模式。描述参考图像,指出唯一动作,描述环境中的微小运动,描述镜头,然后锁定身份并禁止剪辑。

如何在本页使用这些提示词

选择一条配方,点击复制,然后粘贴到本页工作室的提示词框中。如果配方需要照片,请上传一到两张。工具按每张静态图片 1 个积分计费,标准 12 秒视频为 15 个积分。

如果结果接近但不对,一次只改一个分句。同时改三处,你就无法判断是哪一处起了作用。

提示词库

复制一个模板粘贴到上方的工作室。保留「保持」和「避免」条款,只替换服装、颜色或场景。

午夜辫子——哥特肖像

最适合竖版肖像,带有暗黑编辑风格。

Use the uploaded photo as the primary reference. Preserve the subject's recognizable facial identity, facial proportions, skin tone, approximate age and natural anatomy. Change the hairstyle into long violet braids decorated with delicate silver hair accessories. Replace the clothing with an elegant black-and-purple velvet outfit with subtle lace and metallic details. Place the subject in a moonlit Victorian garden with soft fog, glowing lanterns and dark autumn foliage. Use photorealistic editorial photography, realistic fabric textures and cinematic soft lighting. Keep the same number of people. Avoid altered facial identity, extra limbs, duplicated fingers, distorted anatomy, text, logos and watermarks.

宇宙公主——天界长裙

更明亮的外观,同时仍有万圣节气息。

Edit the uploaded photo into an original celestial Halloween princess portrait. Preserve the subject's recognizable face, natural skin texture, body proportions and pose. Replace the outfit with a midnight-blue fantasy gown featuring silver embroidery, softly glowing star patterns and realistic flowing fabric. Add a delicate silver moon crown and luminous accessories. Place the subject inside a moonlit observatory with a starry sky visible through tall windows. Use elegant fantasy fashion photography, cinematic rim lighting, and a deep-blue, silver and violet palette. Keep the person recognizable. Avoid altered facial identity, extra people, distorted anatomy, text or watermarks.

万圣节机器人——可穿戴服装

面部保持人类;只有服装变成机械风格。

Using the uploaded photo as reference, transform the subject into a futuristic Halloween robot costume portrait. Preserve the subject's recognizable face, skin tone, head proportions and original pose. Create a realistic wearable robot costume with metallic-looking armor panels, subtle blue LED accents and articulated shoulder pieces. The outfit should look like a believable costume worn by a real person, not a new mechanical character. Use dramatic blue-and-orange cinematic studio lighting and realistic material reflections. Do not replace the subject's face with a robot head. Avoid extra limbs, distorted hands, duplicated body parts, logos and text.

南瓜宠物——宠物服装

适用于猫和狗;保留品种与斑纹。

Use the uploaded pet photo as the primary reference. Preserve the animal's original breed, fur color, distinctive markings, eye color, facial appearance and natural anatomy. Dress the pet in a soft pumpkin-inspired Halloween costume with a small green leaf accessory. Keep the costume fitted naturally without covering the animal's eyes or changing its species. Place the pet in a cozy autumn photography studio with warm lighting and miniature pumpkins. Create a photorealistic professional pet portrait with realistic fur texture and natural shadows. Avoid extra animals, unnatural paws, duplicated limbs, text and watermarks.

牛仔与天使——情侣肖像

两张照片、两个身份,输出中恰好两个人。

Use the two uploaded photos as identity references. Photo one is the cowboy, photo two is the angel. Preserve both faces exactly: same facial features, skin tone, hair color and body proportions, and do not swap the two people. Dress the first person in a western cowboy costume with a hat, suede vest and boots. Dress the second person in a soft white angel costume with feathered wings and a subtle halo. Place them side by side in a warm autumn barn doorway with hay bales, string lights and pumpkins. Use photorealistic editorial photography, natural expressions, consistent scale and realistic fabric. Keep exactly two people. Avoid a third person, merged faces, extra limbs, text and watermarks.

幽灵肖像——得体妆容

没有血腥的万圣节妆容,所以仍适合用作头像。

Edit the uploaded photo into a cinematic haunted-house portrait for Halloween. Preserve the subject's recognizable face, skin tone and natural anatomy. Add tasteful Halloween makeup: smoked eyes, a subtle stitched scar on one cheek and deep burgundy lips. Replace the background with a dim Victorian parlor: dusty velvet curtains, a candelabra, a gilt mirror and cold moonlight through a tall window. Use soft directional light from the left, realistic skin texture, shallow depth of field and a muted teal-and-amber grade. Keep the same number of people. Avoid horror gore, altered identity, extra people, text and watermarks.

森林女巫——苔藓与烛光

更安全的团体服装选项,不需要道具。

Edit the uploaded photo into a forest witch portrait. Preserve the face, skin tone, age and body proportions. Dress the subject in layered charcoal and moss-green robes with a wide-brim hat, dried flowers and a silver pendant. Place them in a foggy pine forest at blue hour with floating candle lanterns and a faint circle of chalk symbols on the ground. Use cinematic fantasy photography, soft volumetric light and realistic wool and leather textures. Keep exactly one person. Avoid cartoon styling, altered facial identity, extra people, text and watermarks.

夜光恐龙——宠物

2026 年宠物服装趋势,不用给动物上色。

Use the uploaded pet photo as the primary reference. Preserve the animal's breed, fur or skin, markings, eye color and natural anatomy. Dress the pet in a glow-in-the-dark dinosaur costume: a soft green hood with rounded spikes, a small tail and faint luminous seams. Keep the costume comfortable and do not cover the eyes. Place the pet on a dark studio floor with soft UV-green rim light and a few paper bats in the background. Use photorealistic pet photography, realistic fur texture and clean shadows. Avoid extra animals, changed species, glowing eyes, duplicated limbs, text and watermarks.

万圣节夜骑——图像转视频

在静态图片生成后使用;保持动作幅度小。

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic Halloween portrait animation. Keep the subject’s recognizable face, costume, hairstyle and body proportions consistent with the reference. The subject slowly turns their head toward the camera with a subtle mysterious expression. A gentle breeze moves loose hair strands and the edges of the clothing. Soft fog drifts through the moonlit garden. The camera performs a slow and smooth push-in. Maintain realistic motion, stable anatomy and consistent lighting. Do not change the person’s identity, replace the costume, add characters or introduce sudden scene cuts. No text or watermarks.

南瓜地漫步——图像转视频

适合宠物的运动提示词,固定镜头。

Use the provided image as the exact reference. Animate a calm three-second shot of the pet sitting in the pumpkin studio. The pet blinks, flicks one ear and turns its head slightly toward a sound off-camera. A small paper bat drifts past in the foreground. Keep the same breed, fur markings, costume and body proportions. The camera stays locked with a very slow push-in. Warm practical lighting, realistic fur movement, no new animals, no costume changes, no text and no watermarks.

黑色星期五海报——纯文本

无需照片的提示词,适合促销图或故事封面。

Create a premium Black Friday sale poster for an AI video studio. Deep charcoal background with a subtle green glow, one bold cinematic motorcycle silhouette in the lower third, floating price tags and abstract speed lines. Leave a clean empty rectangle in the centre for a headline. Use a limited palette of charcoal, lime green and warm white, crisp geometric shapes, realistic studio lighting and generous negative space. No words, no letters, no logos, no watermarks.

情侣服装——统一剪影

用于两张照片,服装应看起来像同一套。

Use the two uploaded photos as identity references and preserve both faces exactly. Dress both people in a coordinated vampire costume set: matching black velvet capes with deep red lining, high collars and subtle silver clasps. Place them in front of a foggy stone archway with two candelabras. Use a symmetrical composition, both people at the same scale, cinematic red-and-charcoal lighting and realistic fabric. Keep exactly two people and do not swap their identities. Avoid a third person, merged faces, extra limbs, text and watermarks.

使用流程

  1. 1

    复制一条配方

    从已经保护身份并列出失败情况的提示词开始。

  2. 2

    粘贴并调整

    更改服装颜色或场景。保留与避免分句保持原样。

  3. 3

    在本页运行

    每张静态图片 1 个积分。喜欢这一帧后,可用同一提示词生成视频。

常见问题

我可以在 Gemini、ChatGPT 或 Midjourney 中使用这些提示词吗?

可以。它们是普通英文,不依赖本站。不同模型的结果会有差异,但结构——保留、更改、场景、光线、避免——可以迁移。

为什么提示词这么长?

因为约束本身就是产品的一部分。短提示词会把身份、服装细节和人体结构交给运气;长版本会明确写出不能改变的内容。

每条提示词都需要照片吗?

不需要。黑色星期五海报提示词是纯文本。服装和宠物提示词搭配照片效果最好,因为它们是保留身份的图像编辑。

我可以出售用这些提示词生成的图片吗?

本站许可允许商业使用,且输出无水印。如果你计划把图片本身作为商品转售,也请查看模型提供方的条款。

如何把静态图片变成视频?

使用提示词库末尾的图像转视频提示词。保持动作幅度小:转头、眨眼、飘动的雾和缓慢的镜头移动。

先看效果,再决定买

上传照片,选风格,下载图片。1 积分,无水印,可商用。

Make my image