真实可用的预设

可以直接复制的万圣节 AI 提示词

提示词只有在能出结果时才有用。这 8 条就是本站造型背后的提示词,包含让脸保持可认、让服装读得出来的约束。

  • 保留你的脸
  • 新账号 10 积分 = 5 张图
  • 失败退回积分
AI 生成的万圣节肖像:骷髅妆、黑色蕾丝裙和烛光
示例由 gpt-image-2 生成;你的成片会以自己的照片为参考。

每张照片一个人、脸清晰,JPG/PNG/WebP,最大 10 MB。半身或全身照能给走路引擎更多信息。

生成即表示你确认照片属于你,或照片中的人已同意。每张照片一个人(双人造型两张),仅限成年人,不要使用名人照片。

怎么用这些提示词

每条提示词都假设有一张上传的参考照片(双人造型两张)。把它粘到支持参考图的编辑器里;不要当成纯文生图提示词,然后期待同一个人。复制整条,包括末尾的“避免”清单。

真正有用的是约束

服装描述最吸引注意,但约束才让结果可用:保留脸、保持人数、避免多余肢体、不要文字和水印。没有这些句子,提示词会产出更漂亮的混乱,而不是肖像。

先提示词,再预设

如果你就在本站,不需要粘贴任何东西:选对应预设、上传照片即可。提示词展示出来,是为了让你学结构、拿到别的工具里改,或者检查预设到底要求了什么。

提示词解决不了什么

提示词救不回一张脸太小的照片,不能把合影变成单人,也不能保证某个名人的精确长相。它也无法凭空加一个真实品牌 logo 或电影级精确服装。照片和模型决定上限。

提示词库

下面是各预设实际使用的图片提示词。可以复制一条,或打开对应造型页直接生成。

Midnight Braids

打开预设
Use the uploaded photo as the primary reference. Preserve the subject's recognizable facial identity, facial proportions, skin tone, approximate age and natural anatomy. Transform the subject into an original gothic Halloween fashion portrait. Change the hairstyle into long violet braids decorated with delicate silver hair accessories. Replace the clothing with an elegant black-and-purple velvet outfit featuring 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, natural skin details and cinematic soft lighting. Keep the same number of people as the uploaded photo. Avoid altered facial identity, extra limbs, duplicated fingers, distorted anatomy, text, logos and watermarks.

Cosmic Princess

打开预设
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, realistic fabric details, cinematic rim lighting, and a deep-blue, silver and violet colour palette. Keep the person recognizable. Avoid altered facial identity, extra people, distorted anatomy, text or watermarks.

Galactic Hero

打开预设
Using the uploaded photo as reference, transform the subject into an original futuristic Halloween hero costume portrait. Preserve the subject's recognizable face, skin tone, head proportions and original pose. Create a realistic wearable hero costume with sculpted metallic armour panels, subtle blue LED accents and articulated shoulder pieces. The visor is pushed up so the face stays visible. Place the subject on a neon-lit city rooftop at night. Use dramatic blue-and-orange cinematic lighting and realistic material reflections. Keep the person recognizable. Do not replace the face with a helmet. Avoid extra limbs, distorted hands, duplicated body parts, logos and text.

Forest Witch

打开预设
Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, skin tone, body proportions and pose. Transform the subject into an original forest witch portrait: a moss-green velvet cloak with dried flowers and embroidered vines, a weathered leather corset, and an antique lantern held in one hand. Place the subject in a foggy autumn forest with tall trees and soft green light. Use photorealistic fantasy fashion photography, natural skin texture and cinematic mist. Keep the same number of people as the photo. Avoid altered facial identity, extra limbs, distorted hands, text, logos and watermarks.

Skeleton Glam

打开预设
Edit the uploaded photo into an elegant skeleton glam Halloween beauty portrait. Preserve the subject's recognizable facial identity, skin tone, face shape and natural anatomy. Apply skeleton makeup in ivory and charcoal, keeping the eyes, brows and expression readable. Dress the subject in a black lace evening dress with silver earrings. Place the subject in a candlelit ballroom with dark velvet curtains. Use photorealistic high-fashion beauty photography, detailed makeup texture, warm candlelight and cool shadows. Keep the same number of people. Avoid altered facial identity, extra limbs, distorted features, text, logos and watermarks.

Haunted Unicorn

打开预设
Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, skin tone and body proportions. Transform the subject into an original haunted unicorn mascot-inspired Halloween look: a soft pastel-gothic hood with a sculpted unicorn horn, iridescent fabric, dark lace gloves, and a foggy abandoned carnival background with muted pink lights. Use photorealistic costume editorial photography, whimsical but eerie, cinematic fog and soft practical lights. Keep the same number of people. Avoid altered facial identity, extra limbs, distorted hands, text, logos and watermarks.

Pumpkin Pet

打开预设
Use the uploaded pet photo as the primary reference. Preserve the animal's original breed, fur colour, distinctive markings, eye colour, 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.

Cowboy & Angel

打开预设
Use the two uploaded photos as references and create one coordinated Halloween couple portrait. Preserve each person's recognizable facial identity, skin tone, body proportions and natural anatomy. Dress person one as a polished cowboy with a wide-brim hat, leather details and boots. Dress person two as an original angel with layered ivory fabric and softly glowing feather wings. Place them together in a golden-hour desert with warm dust in the air. Use photorealistic editorial fashion photography, consistent lighting across both faces and a coordinated pose. Exactly two people. Do not swap faces, add a third person, distort hands, add text, logos or watermarks.

常见问题

可以拿到 Midjourney 或 DALL·E 里用吗?

可以粘到任何支持参考图的编辑器里,但模型和输入格式会影响结果。纯文生图工具不会像参考图编辑器那样保留你的脸。

在本站需要自己粘提示词吗?

不需要。选预设会自动运行同一条提示词。复制按钮是给想拿到别处用的人准备的。

为什么要有“避免”清单?

它列出肖像最致命的失败模式:脸变了、多手多脚、手指重复、文字和水印。把它们写出来能降低出现概率。

可以改提示词里的颜色吗?

可以。保留“保留”和“避免”部分,改服装和场景描述。那是最安全的实验位置。

示例图是用这些提示词生成的吗?

造型页的示例图用 gpt-image-2 生成。它们是视觉参考,不保证换一个模型或照片会得到同样结果。