30多个适用于照片、艺术、广告和设计的最佳 ChatGPT 图像提示词
- 1. 为什么结构化的 ChatGPT 图像提示词效果更好
- 2. 弱提示词与可用于生产的提示词对比
- 3. ChatGPT 图像提示词框架
- 4. 如何编写更好的 ChatGPT 图像提示词
- 5. 32 个最佳 ChatGPT 图像提示词
- 6. 商业类别及如何进行调整
- 7. 可重复使用的 ChatGPT 图像提示词模板
- 8. 关于 ChatGPT 图像提示词的常见问题
- 9. 结论
最好的 ChatGPT 图像提示词(prompt)绝不仅仅只是说出一个主体。它们会解释视觉风格、构图、光照、色彩、材质、文字以及需要保持一致的细节。这种结构有助于将一个粗略的想法转化为更容易用于头像、产品宣传、海报、社论插图或社交媒体帖子的图像。
本指南收集了 32 个可直接复制的 ChatGPT 图像提示词,涵盖肖像、时尚、旅行、产品、海报、插画和创意实验。前 14 个改编自 GPT Image 2 提示词 页面中无竞争的精选案例。其余 18 个则使用了来自 GoEnhance 图像提示词库的其他公开案例图像。每个条目都包含一张对应的案例图像;图像下方的提示词是实际的改编版本,并非宣称它是源页面的精确措辞。
如需获得更广泛的灵感,请浏览更广阔的图像提示词库,并根据您的主题、受众和输出格式调整结构。
1. 为什么结构化的 ChatGPT 图像提示词效果更好
“制作一张美丽的时尚肖像”给了 ChatGPT 一个方向,但它让绝大多数的视觉决定处于开放状态。模型仍然必须猜测主体的姿势、服装、背景、拍摄距离、光线方向、调色板和图像比例。这就是为什么一个简单的请求可以产生一张引人注目但并不符合真实设计简报的图像。
一个有用的图像提示词就像一份紧凑的艺术指导说明。它告诉模型应该优先考虑什么,应该避免什么。你不需要描述每一个可能的细节,但你应该定义视觉层级:首先是主体,其次是环境,最后是收尾细节。
2. 弱提示词与可用于生产的提示词对比
弱提示词
Create a stylish woman in a beautiful garden.
可用于生产的提示词
Create a photorealistic editorial portrait of a young woman standing in a curated garden at late afternoon. She wears a cream linen dress, small gold earrings, and natural makeup. Frame her from the waist up at eye level, with flowering branches creating a soft foreground frame. Use warm directional sunlight, gentle shadows, muted sage and peach tones, shallow depth of field, realistic skin texture, natural hands, and an 85mm portrait lens. Leave clean negative space in the upper-right corner for a headline. Vertical 4:5 composition, no watermark, no extra text, no distorted features.
第二个提示词为模型提供了主体、背景、构图、光照计划、使用要求和负面约束。它更容易生成、审查和修改。
3. ChatGPT 图像提示词框架
将此结构作为起点:
[style/type], [subject], [action or pose], [scene], [composition], [lighting], [color and texture], [text instructions], [aspect ratio/use], [negative constraints]
产品提示词可能以材质和拍摄角度开头,而肖像提示词可能以主体的身份和表情开头。顺序可以改变,但提示词应该回答以下问题:
- 主体是什么?
- 观众应该首先注意到什么?
- 主体在哪里,画面是如何构图的?
- 什么光线、颜色和质感定义了这幅图像?
- 需要什么文字、比例或设计空间?
- 模型应该避免改变或凭空捏造什么?
4. 如何编写更好的 ChatGPT 图像提示词
4.1 从图像类型开始
使用清晰的类型,例如写实肖像、产品广告、社论插图、旅行海报、扁平矢量双联画、水彩童书艺术、电影剧照或极简品牌图形。在你添加细节之前,类型就已经设定了视觉预期。
4.2 描述主体与姿势
指定主体的外观、表情、服装、配饰以及在画面的位置。如果图像基于参考图,请说明身份、面部特征、发型和比例应保持一致。对于产品,请描述其形状、材质、表面处理和标签位置。
4.3 围绕主体构建场景
当场景包含地点、一天中的时间、背景元素和深度线索时,它就更容易生成。“海岸线上方的阳台”比“好看的背景”更有用。当这些细节支持你的概念时,可以加入玻璃栏杆、盆栽树、海洋地平线或建筑纹理。
4.4 指导构图与镜头语言
使用诸如极特写、半身肖像、全身构图、居中对称、四分之三侧面、俯视平铺(flat-lay)、低角度、浅景深、引导线和负空间等术语。只有在能改善预期外观时,才添加镜头或摄影风格。
4.5 控制光照、调色板和质感
光照可以营造氛围并将一张图像与另一张区分开来。尝试柔和的窗光、正午的硬直射光、黄金时段的逆光、低调的暗调照明(low-key)、霓虹边缘光或漫射的阴天光。然后加入有限的调色板和物理质感,如胶片颗粒、湿玻璃、纸纤维、织物纹理、金属反光或水彩颜料。
4.6 为文字提供精确的指令
如果图像需要标题,请将确切的文本写在引号中,并指定其位置、大小、字重和对比度。如果文字准确性并非必不可少,请请求干净的负空间,稍后在设计工具中添加排版。这通常能让你对最终布局拥有更多控制权。
为了便于访问发布,请将重要的图像内容与实用的替代文本(alt text)配对,并保持图像内的文字简短。有关一般网页发布指南,请参阅 W3C 图像教程和 Google 的图像 SEO 指南。
4.7 添加有用的负面约束
负面约束应该针对常见的失败情况:无水印、无多余手指、无重复对象、无畸变文字、无塑料感皮肤、无随机徽标、无裁剪主体、无杂乱背景,并且不改变所提供的产品形状。保持列表的相关性,而不是添加所有可能的警告。
5. 32 个最佳 ChatGPT 图像提示词
每个示例都包含一个实际用例、一张匹配的公共案例图像和一个可直接复制的提示词。请替换括号中的字段,或将场景调整为您自己的品牌、受众和格式。
5.1 精选 GPT Image 2 案例改编
1. 局部色彩琥珀色墨镜肖像
用例: 奢侈眼镜、时尚社论和个人资料形象。

Create a striking studio portrait of a confident man facing the camera, wearing translucent amber-orange oval sunglasses with ice-gradient lenses. Render the portrait mostly in black and white while keeping only the sunglasses in vivid orange. Dress him in a black turtleneck against a minimalist light-grey background. Use an 85mm lens, f/2, soft directional studio lighting, realistic skin texture, crisp eyewear reflections, centered head-and-shoulders framing, editorial luxury mood, vertical 4:5 composition, no logo, no watermark, no extra text.
2. 超现实哥特恐怖肖像
用例: 专辑封面、恐怖小说封面和黑暗社论宣传活动。

Create a cinematic surreal portrait of a pale young woman standing in a cosmic void. Surround her with elegant pale hands tipped with long crimson nails and a pale serpentine creature curling through the background. Add a glowing turquoise-eyed atmosphere and a deep red moon behind her. Use dark charcoal, dusty beige, muted teal, faded blue, and deep crimson, with low-key lighting, controlled negative space, realistic anatomy, dramatic 4:5 vertical framing, and no gore, logos, watermark, or random text.
3. 南亚庭院时尚肖像
用例: 传统时尚、文化叙事和生活方式宣传活动。

Create a photorealistic portrait of a stylish South Asian woman standing in a traditional courtyard filled with bougainvillea. She wears a deep-crimson embroidered outfit, with a flowering branch framing the top of the image. Include warm yellow walls, turquoise doors, textured stone, and soft natural daylight. Use an 85mm portrait lens, f/2, waist-up composition, realistic hands and anatomy, rich textile detail, gentle shadows, warm editorial color grading, vertical 3:4, no modern logos, no watermark, no extra text.
4. 舒适的深夜卧室自拍
用例: 游戏创作者、生活方式博客和社交资料内容。

Create a photorealistic vertical late-night bedroom selfie of a young East Asian woman relaxing in a desk chair beside a gaming and study setup. She wears a white camisole with a small cute graphic and soft pink pajama pants. Include dual monitors, a warm desk lamp, an iced coffee, instant noodles, a bed, a cat pillow, monochrome posters, and a city window at night. Use monitor glow and warm ambient light, natural skin texture, handheld smartphone framing, lived-in details, subtle grain, realistic proportions, 4:5 vertical, no readable random text or watermark.
5. 山涧旅行肖像
用例: 旅行摄影、户外品牌和夏季社交宣传活动。

Create an authentic smartphone travel photograph of a young East Asian woman sitting in a clear shallow mountain stream on a bright summer day. Her dark hair is naturally wet, and she wears a loose light-grey sweatshirt with casual shorts. Show transparent water, realistic ripples, smooth rocks, lush forest vegetation, sparkling droplets, and dappled sunlight across her face. Use a slightly low camera angle close to the water surface, natural wide-angle perspective, vibrant but realistic greens and blues, peaceful mood, vertical 4:5, no studio lighting, no plastic skin, no watermark.
6. 祖母绿与黄金奢华肖像
Context: 高级个人品牌、男装和杂志封面。

Create a cinematic close-up portrait of a stylish middle-aged man with thick naturally curly dark hair and a neatly trimmed salt-and-pepper beard. He wears round amber-brown sunglasses and a black oversized hoodie. Add a rich emerald rim light on the right side of his face and a soft warm-gold key light from the front. Use a smoky emerald-to-black studio background, realistic beard hairs, natural facial imperfections, shallow depth of field, 85mm lens, f/1.8, centered 3:4 composition, no logo, no watermark, no extra text.
7. 更紧凑的半脸恐怖特写
用例: 恐怖海报、游戏概念艺术和戏剧性缩略图。

Create an extreme close-up split-face horror portrait cropped from the hairline to the chin. The left half is a pale human face with blunt bangs and a sharp dark eye; the right half transforms into a monstrous jaw with jagged teeth and a black void mouth. Add fine red glowing cracks along the transformation seam and one blood-red catchlight. Use a pure black background, high-contrast single-source rim lighting, hyper-detailed skin and teeth, direct eye contact, 9:16 vertical composition, no gore splashes, no watermark, no random text.
8. 动感红色背景时尚肖像
用例: 街头服饰宣传、实验时尚和动感海报。

Create a photorealistic full-body fashion portrait of a stylish young man standing against a seamless vivid-red studio background. He wears an oversized cream ribbed sweatshirt, distressed denim shorts, black socks, chunky sneakers, a simple necklace, and dark futuristic sunglasses. Stretch the left side of his clothing into translucent textile fibers and motion trails while keeping his face, crossed arms, legs, and shoes sharp. Use strong studio light, realistic fabric texture, dramatic shadows, 35mm fashion photography, analog grain, vertical 4:5, no distorted anatomy, no logos, no extra text.
9. 水晶鳞片奇幻美妆特写
用例: 美妆宣传、奇幻肖像和社交资料艺术。

Create an ultra-realistic extreme close-up beauty portrait of a young East Asian woman with luminous skin, warm brown eyes, coral-pink lips, loose black waves, and dewy makeup. Blend translucent iridescent crystal-scale texture naturally into both cheeks, with fine holographic glitter across the face. Use cyan-blue key light from the right, warm pink fill from the left, pastel bokeh, volumetric glow, sparkling particles, translucent pink tulle, realistic pores, 85mm macro lens, f/1.4, editorial luxury mood, no text, no watermark.
10. 海滨静谧奢华肖像
用例: 度假时尚、旅游杂志和生活方式宣传。

Create an ultra-realistic full-body fashion portrait of a young man on a modern oceanfront balcony in bright morning sunlight. He wears an olive-green linen shirt, oversized black trousers, white minimalist sneakers, and black sunglasses. One hand holds a takeaway coffee while the other rests in his pocket. Include cream architecture, polished marble, a glass railing, a potted olive tree, deep-blue ocean, sandy beach, and clear sky. Use relaxed quiet-luxury styling, natural shadows, 85mm lens, f/2.8, realistic reflections, vertical 4:5, no logo or watermark.
11. 经典低趴车(Lowrider)电影感肖像
用例: 汽车文化、音乐封面和街头服饰品牌。

Create a cinematic golden-hour portrait of a tattooed man seated on the hood of a cherry-red 1960s lowrider. Show a white ribbed tank top, tan cargo pants, detailed tattoos, gleaming chrome, reflective candy paint, a dusty asphalt lot, a chain-link fence, and a hazy downtown skyline. Use strong leading lines from the car and fence, shallow depth of field, warm directional light, natural shadows, rich 35mm film grain, realistic skin and fabric, balanced composition, no weapons, no random logos, no extra text.
12. 社论扁平矢量双联画
用例: 杂志封面、创意个人资料和品牌形象系统。

Create a cohesive diptych-style editorial illustration with two close-up portraits in the same modern flat-vector fashion aesthetic. Give the woman flowing dark hair, gold hoops, layered necklaces, and a white sleeveless top; give the man textured platinum-blond hair and an oversized white crew-neck shirt. Use geometric facial planes, crisp linework, subtle tonal shading, a warm tan panel behind the woman, a bold red panel behind the man, and a shared ivory canvas. Luxury magazine style, clean color blocking, vertical 4:5, no photorealism, no 3D rendering, no text, no logos.
13. 韩国网络漫画风格旅行家庭
用例: 旅游业、家庭旅行和插画明信片。

Create a soft minimalist Korean webtoon illustration of a traveling family of two parents and two children exploring [BACKGROUND]. Use cute simplified faces, thin clean line art, muted beige, sage green, dusty blue, cream, and light pink, rounded proportions, flat shading, and a cozy travel-journal mood. The father carries a backpack and folded map; the mother carries a camera or shoulder bag. Use full-body vertical 4:5 composition, family in the foreground, open space around them, balanced landmarks, optional location title at the top, no clutter, no random text.
14. 琥珀色墨镜影棚变体
用例: 眼镜产品叙事和单色宣传概念。

Create a refined head-and-shoulders studio portrait of a mature man with curly dark hair and a short salt-and-pepper beard. Keep the portrait nearly monochrome, but make the round amber sunglasses the single saturated color accent. Use a dark blue-grey background, soft frontal light, a subtle cool rim light, crisp lens reflections, natural skin texture, and a quiet luxury editorial mood. Center the subject with clean negative space, 85mm portrait lens, f/1.8, vertical 3:4, no logo, no watermark, no extra text.
5.2 额外的 GoEnhance 公开案例改编
15. 严格遵守提示词的城市构图
用例: 房地产概念、城市海报和版式测试。

Create a clean photorealistic city composition designed to test precise layout adherence. Place a recognizable skyline in the middle distance, a broad empty foreground, and clearly separated architectural blocks on the left and right. Keep the horizon level, preserve generous negative space in the upper third, and use soft daylight, realistic scale, crisp edges, and restrained neutral colors. Use a wide 16:9 composition, no tilted perspective, no duplicated buildings, no random signs, no watermark.
16. 专业医疗证件照
用例: 诊所网站、医疗保健个人资料和教育内容。

Create a professional photorealistic headshot of a friendly doctor specializing in [SPECIALTY]. Use a clean neutral background, a white coat over understated professional clothing, natural eye contact, relaxed shoulders, and a credible approachable expression. Use soft key light, subtle fill, realistic skin texture, accurate hands if visible, 85mm portrait perspective, centered waist-up framing, clean healthcare branding, vertical 4:5, no medical claims, no readable patient information, no logo, no watermark.
17. 霓虹传送门动作场景
用例: 游戏主视觉图、电影概念和动作缩略图。

Transform a supplied portrait into a hyper-realistic cinematic action scene on a rain-slicked urban street. Place a glowing dimensional neon portal behind the subject, with cyan and magenta light reflecting across wet pavement and nearby architecture. Keep the subject’s identity, face, clothing, and proportions consistent. Use a low-angle 35mm composition, controlled motion blur, atmospheric rain, strong rim lighting, realistic reflections, high contrast, 16:9 landscape, no extra limbs, no random text, no watermark.
18. 荷塘清晨场景
用例: 冥想艺术品、养生内容和自然海报。

Create a serene photorealistic lotus pond at early morning. Place several pink and white lotus flowers above broad green leaves, with delicate mist hovering over still water and soft mountains fading into the background. Use gentle sunrise light, pastel peach and sage tones, realistic water reflections, fine petal texture, balanced negative space, and a calm contemplative atmosphere. Use a vertical 4:5 composition, no people, no text, no artificial neon colors, no watermark.
19. 修复肖像转换
用例: 全家福修复、纪念项目和个人档案。

Restore the supplied vintage portrait while preserving the person’s identity, facial structure, expression, clothing, and original pose. Remove scratches, dust, fading, and color damage; rebuild missing detail conservatively and improve contrast with a natural film palette. If the brief requests a hairstyle transformation, change only the hair while keeping the face and age consistent. Use realistic skin texture, period-appropriate color, gentle sharpening, and a clean archival finish. No beauty-filter effect, no new accessories, no altered facial features, no watermark.
20. 刺绣与圆珠笔素描
用例: 工艺品牌、艺术教育和混合媒介社论艺术。

Create a mixed-media portrait combining fine blue ballpoint-pen hatching with raised embroidery details. Use a simple central subject, visible paper fibers, stitched floral accents, delicate thread shadows, and carefully controlled line density. Keep the composition editorial and uncluttered, with a warm cream paper background and restrained indigo, coral, and mustard accents. Show the physical material contrast clearly, use vertical 4:5 framing, no photorealistic skin, no random lettering, no watermark.
21. 角色替换编辑
用例: 分镜头脚本、游戏概念开发和角色设计。

Edit the supplied scene by replacing only the main character with [NEW CHARACTER]. Preserve the original camera angle, pose, lighting direction, environment, perspective, shadows, and all unchanged background objects. Match the new character’s scale, contact shadows, clothing texture, and color grade to the scene. Keep the same framing and aspect ratio. Do not alter the architecture, props, sky, or composition; no extra characters, logos, text, or watermark.
22. 传达情感的肖像
用例: 小说封面、宣传叙事和社论肖像。

Create a cinematic portrait that communicates [EMOTION] through posture, eye direction, facial tension, hand placement, and surrounding light rather than exaggerated expressions. Place the subject in a simple atmospheric environment with one meaningful prop, soft directional lighting, subtle color contrast, and realistic material detail. Use a medium close-up, shallow depth of field, natural anatomy, quiet visual hierarchy, and a 4:5 vertical frame. Avoid melodrama, plastic skin, duplicated objects, random text, logos, and watermark.
23. 宝石色调马赛克森林
用例: 墙面艺术、文化插画和创意背景。

Create a highly detailed fine-art forest avenue rendered as a dreamlike mosaic painting. Use jewel-toned emerald, sapphire, ruby, and amber fragments to build trees, foliage, a winding path, and atmospheric depth. Keep the scene readable from foreground to background, with rhythmic tile shapes, luminous highlights, and a sophisticated gallery-art finish. Use a wide 16:9 composition, controlled symmetry, no photorealistic people, no random text, no logo, no watermark.
24. 逼真的史前动物
用例: 博物馆教育、自然历史插画和科学叙事。

Create a scientifically grounded wildlife scene featuring a realistic Platybelodon in a late-Miocene grassland. Show the animal’s shovel-like lower jaw, broad body, believable skin folds, dust on the feet, and natural herd-scale environment. Use warm low-angle sunlight, a restrained earth-tone palette, atmospheric distance, accurate anatomy, and documentary wildlife photography framing. Use landscape 16:9 composition, no fantasy armor, no modern objects, no labels, no watermark.
25. 珠宝与服装编辑
用例: 时尚目录、试穿概念和产品造型。

Edit the supplied fashion portrait by changing only the jewelry and clothing. Replace the outfit with [CLOTHING DESCRIPTION] and add [JEWELRY DESCRIPTION], matching the original body shape, pose, face, hairstyle, camera angle, background, shadows, and color grade. Make the fabric folds, metal reflections, contact shadows, and scale physically believable. Keep the original identity unchanged, preserve the composition, use vertical 4:5 framing, and include no extra accessories, logos, text, or watermark.
26. 汽车与背景转换
用例: 汽车房源列表、概念板和广告样机。

Replace only the background of the supplied car image with [NEW LOCATION]. Preserve the exact vehicle shape, paint color, wheels, windows, reflections, camera angle, and proportions. Match the new environment’s perspective, light direction, contact shadow, atmospheric haze, and road texture to the vehicle. Keep the car sharp and central, use a premium automotive-advertising look, landscape 16:9 composition, no model changes, no logo edits, no added text, no watermark.
27. 极简主义结婚肖像
用例: 婚礼网站、请柬和情侣公告。

Create an extremely realistic minimalist wedding portrait of a couple standing close together against a warm neutral background. Use elegant understated clothing, natural expressions, relaxed hands, soft window light, gentle shadows, and a quiet intimate atmosphere. Keep the composition uncluttered with generous negative space for later invitation text. Use realistic fabric, skin, and ring details, medium portrait framing, vertical 4:5, no extra jewelry, no distorted hands, no random text, no watermark.
28. 教育类 AI 艺术拼贴
用例: 课堂幻灯片、博物馆教育和视觉解说。

Create an educational visual collage explaining [TOPIC]. Arrange the main subject in the center and surround it with clearly separated supporting scenes, objects, and simple visual labels. Use a consistent illustration style, limited color palette, strong hierarchy, generous spacing, and accurate relationships between the elements. Keep text short and render the exact labels “[LABEL 1]” and “[LABEL 2]” only if needed. Use a clean 16:9 layout, no decorative clutter, no false data, no watermark.
29. 城市街道上的巨型人物
用例: 超现实海报、电影概念和比例对比艺术。

Create a cinematic wide-angle city street scene with a colossal person seated calmly among ordinary buildings and traffic lanes. Make the scale relationship clear through tiny architecture, pedestrians, street markings, and perspective lines while keeping the figure’s anatomy believable. Use overcast daylight, realistic atmospheric depth, restrained urban colors, subtle reflections, and a low camera angle. Landscape 16:9 composition, no destruction, no panic, no random signage, no extra limbs, no watermark.
30. 艺术肖像编辑
用例: 美容改造、社交头像和社论实验。

Transform the supplied portrait into a refined artistic editorial image while preserving the subject’s identity, face shape, expression, hairstyle, and pose. Apply [STYLE] through controlled brushwork, paper texture, color layering, and background treatment rather than changing the person. Keep facial features sharp enough to remain recognizable, use a limited palette, balanced negative space, and vertical 4:5 composition. Change only the visual medium and background; no extra accessories, distorted anatomy, random text, or watermark.
31. 浪漫蓝墨水肖像
用例: 文具、书籍封面、浪漫宣传活动和墙面艺术。

Create a romantic portrait illustration built from expressive cobalt-blue ink flowing around the subject like soft ribbons. Keep the face and pose readable while allowing translucent ink blooms, feathered edges, and delicate splashes to shape the background. Use a cream paper ground, gentle side lighting, subtle paper grain, elegant negative space, and a calm poetic mood. Vertical 4:5 composition, no black outlines that overpower the face, no random lettering, no logo, no watermark.
32. 极简主义婚礼摄影肖像
用例: 仪式公告、婚礼博客和高级请柬艺术。

Create a premium minimalist wedding photograph of a couple in a bright neutral interior with soft architectural lines. Pose them naturally with subtle eye contact and relaxed hands, wearing timeless formal clothing with realistic fabric texture. Use diffused window light, gentle beige and ivory tones, shallow depth of field, quiet editorial styling, and clean negative space on one side for invitation copy. Vertical 4:5 framing, accurate anatomy, no extra fingers, no random text, no logos, no watermark.
6. 商业类别及如何进行调整
肖像与个人品牌
锁定面部、表情、服装、拍摄距离和光线方向。对于可重复使用的个人资料图像,保持背景简单,并留出一些负空间以放置姓名或职位。
产品与包装概念
描述产品的确切形状、表面处理、标签区域以及与表面的关系。添加拍摄角度和光线反射,然后要求不要有多余的徽标或虚构的产品功能。
海报与社交图形
在进行装饰之前先定义层级。说明确切的标题、文本位置、纵横比和留白。如果拼写至关重要,请考虑在生成后添加最终文本。
社论插画
说明媒介、线条质量、调色板、质感和真实感水平。对于双联画或拼贴画,请为每个面板重复样式要求,以便各个部分感觉像是一个统一的设计资产。
图像编辑与修复
使用清晰的不变量:“仅更改背景”、“保留人脸”或“保持产品形状不变”。这减少了不需要的编辑,并使每一次修改都更容易评估。
7. 可重复使用的 ChatGPT 图像提示词模板
写实肖像模板
Create a [image type] of [subject] in [location]. The subject wears [clothing/accessories] and has [appearance/expression]. Frame the image as [composition] from [camera angle]. Use [lighting], [palette], and [texture]. Intended use: [use case]. [Aspect ratio]. Preserve [identity/invariants]. Avoid [negative constraints].
产品宣传模板
Create a premium product photograph of [product] on [surface] in [environment]. Show [material, finish, and key details] clearly. Use [camera angle], [lens/look], and [lighting]. Leave [location] empty for copy. Keep the product shape and label placement unchanged. [Aspect ratio]. No invented logos, distorted text, extra products, or watermark.
海报模板
Create a [style/medium] poster about [topic]. Place [main subject] in [scene] with [supporting elements]. Use [palette], [lighting], [texture], and [composition]. Add the exact text “[HEADLINE]” at [location] and “[SUBHEAD]” at [location]. Keep the hierarchy clear, [aspect ratio], no misspelled text, random letters, logos, or watermark.
图像编辑模板
Edit the supplied image by changing only [target element]. Preserve [identity, pose, product, camera angle, lighting, background, and proportions]. Match [perspective, shadows, reflections, and texture] to the original. Keep the same framing and aspect ratio. Do not alter [protected elements]. No extra objects, text, logos, or watermark.
8. 关于 ChatGPT 图像提示词的常见问题
ChatGPT 图像提示词应该有多长?
提示词的长度应该足以消除重大的歧义,而不是为了长而长。从主体和预期的图像类型开始,然后添加场景、构图、光照、材质、文本要求,以及仅那些至关重要的负面约束。
我应该包含相机设置吗?
当你想要摄影效果时,相机术语的作用最大。镜头、取景、景深和光线方向可以指导视觉语言,但它们应该支持简报而不是取代简报。
如何改进 AI 生成图像中的文字?
将确切的文本写在引号中,指定它应该出现的位置,并保持简短。对于徽标、包装标签和长标题,请生成带有负空间的干净布局,并在准确性至关重要时在设计工具中添加最终排版。
我应该使用参考图像吗?
当身份、产品形状、姿势或特定的视觉风格需要保持一致时,请使用参考图像。说明应该保留什么以及可以改变什么。不变量越具体,审查结果就越容易。
9. 结论
最可靠的 ChatGPT 图像提示词将创意指导与实际约束结合在一起。从主体开始,定义视觉层级,控制场景和光线,指定格式,最后以一个集中的避免列表收尾。上面的 32 个示例展示了相同的结构如何支持肖像、产品、插画、海报、修复和商业宣传活动。
将每个提示词视为初始简报,而不是固定的咒语。在保持核心结构完整的同时,替换主体、位置、调色板、纵横比和预期的用途。这为您提供了一个可重复的提示词系统,而不是一堆互不关联的单次创意集合。



