AI Image Prompts · 绘画提示词库
出处与授权
本站只收录项目元数据与简介,不对作者正文做转载;完整内容请到上方出处查看。商用前请自行核对原仓库的 LICENSE 与 NOTICE。
展开许可证全文
MIT License Copyright (c) 2026 YouMind-OpenLab Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
LICENSE · 2026-09-26 23:14:14
产品介绍
Recommend curated prompts from a 10,000+ real-world image generation prompt library. Works with ANY AI image model — Nano Banana Pro, Nano Banana 2, Seedream 5.0, GPT Image 1.5, Midjourney, DALL-E 3, Flux, Stable Diffusion, and more. Use this skill when users want to: - Find proven image generation prompts (any model) - Get prompt inspiration for portraits, products, social media, posters, etc. - Create illustrations for articles, videos, podcasts, or marketing content - Browse categorized prompt templates with sample images - Translate and understand prompt techniques
资源信息
许可证信息摘自官方源,可能过时或缺失。商用前请以原仓库的 LICENSE 文件为准。
安装方式
npx skills i YouMind-OpenLab/ai-image-prompts-skill
功能说明
- 按用途分成 11 类,不是随机堆砌
- 全文已入库,站内直接检索、可复制
- 语义检索(描述需求即可)
- 内容改写模式(贴正文生成定制提示词)
- 数据每日同步两次
- MIT 许可,无需 API Key
提示词库 · 2.3万
数据源:社区提示词库(MIT 许可,每日同步)共 15141 条提示词,按 11 个用途归类;同一条可归入多个用途,因此分类条数相加会大于去重后的总数。
命中 2105 条
Extreme Close-up Portrait Prompt
profile-avatarA highly detailed prompt for generating an extreme close-up, photorealistic portrait of a young woman with a short brown bob haircut. The prompt specifies lighting, makeup, expression, and technical details like 8K resolution and hyper-detailed skin texture, suitable for professional portrait photography.
展开提示词全文
Extreme close-up portrait of a young woman facing the camera directly. She has a short brown bob haircut with a precise center part, smooth and neatly styled. Her eyes are large and bright blue, with natural reflections that enhance their depth and clarity. She is wearing soft, natural makeup with subtle foundation, lightly flushed cheeks, softly defined brows, a touch of mascara, and a nude-toned lip color. Small, elegant pearl earrings add a delicate and refined touch to her appearance. Her expression is innocent, calm, and slightly curious, with relaxed facial muscles and a gentle gaze. Her skin texture is highly detailed and realistic, showing natural pores, soft highlights, and smooth tonal transitions. The lighting is professional studio lighting, soft and diffused, evenly illuminating her face without harsh shadows, creating gentle highlights on the cheekbones and subtle shadows under the jawline. The background is clean and minimalistic, light gray in color, softly blurred to keep full focus on the subject. The composition is centered and balanced, with sharp focus on the eyes and shallow depth of field. The image is photorealistic, ultra-high resolution, hyper-detailed skin texture, cinematic quality, HDR effect, sharp focus, 8K resolution, realistic color grading, natural skin tones, and professional portrait photography style.
Passport Photo Generator Prompt (JSON)
profile-avatarA structured JSON prompt designed to generate a grid of six identical, ultra-realistic passport-size photos based on an uploaded image, ensuring strict adherence to official standards for expression, lighting, background, and composition.
展开提示词全文
{
"description": "Ultra-realistic passport-size photo of me (exact face likeness, 99% match, no facial changes).",
"style": "Professional studio portrait suitable for official ID or passport use.",
"outfit": "as per uploaded pic",
"expression": "Neutral closed-lip smile, straight look, both ears visible, shoulders included.",
"lighting": "Even balanced front lighting, no harsh shadows, soft studio tone.",
"background": "Plain light cream or light blue background, clean anduniform.",
"composition": "Centered head and shoulder framing with natural proportions as per passport standards.",
"output": "Generate 6 identical passp ort/ticket-size photos (3x2 grid layout vertical) on one sheet, evenly spaced, print-ready, high-resolution studio quality."
}
Fine-Art Cinematic Monochrome Studio Portrait Prompt (Reference Match)
profile-avatarA highly detailed JSON prompt for generating a fine-art cinematic monochrome studio portrait of a mature confident man, with strict requirements for precisely matching an uploaded reference image's identity, facial structure, beard, and hairstyle, focusing on low-key lighting and high contrast tonal control.
展开提示词全文
{
"master_prompt": {
"scene_type": "fine-art cinematic monochrome studio portrait",
"reference_requirement": {
"identity_match": "precisely match the uploaded reference image",
"facial_structure_accuracy": [
"bone structure",
"eye shape",
"nose proportions",
"jawline definition",
"facial symmetry",
"hairline",
"beard density if present"
],
"skin_texture": "authentic natural skin texture with visible pores and fine lines",
"no_artificial_smoothing": true
},
"subject": {
"type": "mature confident man",
"expression": {
"mood": "serious, contemplative, quietly authoritative",
"gaze": "deep-set eyes looking directly into the camera",
"emotion": "focused and introspective, not aggressive"
}
},
"facial_hair": {
"adaptive": true,
"rules": [
"replicate beard exactly if present",
"match density, length, shape, and edges",
"preserve natural tonal variation including salt-and-pepper if visible",
"beard lines clean and structured"
],
"clean_shaven_rule": "if clean-shaven, emphasize natural jawline definition and skin texture"
},
"hairstyle": {
"adaptive": true,
"requirement": "match hairstyle from reference image exactly",
"details": [
"same volume",
"same direction",
"same hairline",
"natural strand texture",
"subtle specular highlights from overhead lighting"
]
},
"pose": {
"style": "classic thinker pose",
"description": [
"one hand raised to chin",
"thumb along the jawline",
"index finger slightly curled",
"knuckles and tendons subtly defined",
"hand in sharp focus with natural skin texture"
]
},
"lighting": {
"style": "low-key fine-art studio lighting",
"contrast": "high contrast but refined tonal control",
"key_light": {
"position": "slightly above eye level",
"direction": "toward viewer's left angled downward",
"type": "large softbox or diffused beauty dish",
"effect": "strong top-plane highlights with crisp edge definition"
},
"highlight_focus": [
"forehead center",
"nose bridge",
"upper cheekbones",
"beard texture",
"top planes of the hand"
],
"shadow_control": [
"subtle shadow shaping under brow ridge",
"defined cheekbone shadow",
"deep but detailed shadow side",
"midtone detail preserved"
],
"fill_light": {
"position": "minimal fill from front-center",
"purpose": "maintain facial symmetry while preserving depth"
},
"rim_light": {
"intensity": "very subtle",
"location": "hair edges and outer shoulder",
"purpose": "separate subject from black"
}
}
}
}
3D Cartoon Avatar Character Grid Prompt (Pixar Style)
profile-avatarA prompt for generating a 3x3 grid of the same male 3D cartoon avatar character, rendered in Pixar style, showing nine different emotions and expressions, with colorful gradient backgrounds.
展开提示词全文
3D cartoon avatar character showing different emotions arranged in a 3x3 grid, same male character with brown hair, Pixar-style animation, each square showing a different expression (happy, shocked, angry, caring with a puppy, confident, thumbs up, sad, thinking, professional), colorful gradient backgrounds for each panel, soft studio lighting, smooth 3D rendering, expressive eyes and facial expressions, modern emoji-style character design, clean composition, ultra-detailed, vibrant colors, 8K.
Nano Banana Prompt for Art Toy Generation
profile-avatarA Nano Banana Prompt instruction to generate an Art Toy (Lego, Funko, Playmobil, or chosen figure) version of the user, based on a photo of themselves, ensuring the outfit is accurately replicated.
展开提示词全文
You with your Lego, Funko, Playmobil self or the figure you choose wherever you want
Editorial Portrait Prompt for Nano Banana Pro
profile-avatarA highly detailed and structured prompt designed for Nano Banana Pro to generate an ultra-detailed, photorealistic editorial portrait. It specifies aspect ratio, resolution, source image preservation, wardrobe, expression, composition, background, lighting, mood, and technical aesthetic.
展开提示词全文
Aspect ratio 4:5, 2K resolution, ultra-detailed photorealistic editorial portrait. Source: preserve the exact facial features, skin tone, hairstyle, and natural resemblance of the attached reference photo. The subject can be a man or a woman. Wardrobe: Premium black piqué polo shirt, fitted, sharp, with visible cotton piqué texture, deep matte black color. Expression: serious, intense, focused, looking slightly above the camera, not directly at the lens. Composition: Medium close-up (face and shoulders), low-angle shot to convey power and dominance. Background: vibrant red-orange gradient, smooth and intense, no patterns. Lighting: Dramatic and colorful studio lighting. Key light: strong, directional, deep shadows emphasizing facial structure (chiaroscuro). Rim light: strong edge light separating the subject from the background. Mood: intense, mysterious, contrasted. Technique: sharp focus on the face, smooth background gradient, high-quality photorealistic texture, 2K ultra-detail. Apply consistently regardless of gender. Cinematic studio aesthetic, editorial heroic portrait ambiance.
Hyper-Realistic Mirror Selfie Prompt (Pajamas)
profile-avatarA structured JSON prompt for generating a hyper-realistic mirror selfie of a young East Asian woman in plaid pajamas and a camisole, focusing on authentic social media aesthetic, specific clothing details (Calvin Klein waistband), and warm indoor lighting.
展开提示词全文
{
"subject": {
"description": "A beautiful young East Asian woman taking a mirror selfie",
"features": "Long straight black hair, soft facial features, subtle makeup, neutral expression looking slightly away from the camera",
"clothing": [
"White and navy blue thin-striped ribbed camisole with a small bow detail",
"Green and blue plaid pajama pants",
"Visible white Calvin Klein waistband"
]
},
"setting": {
"location": "Modern indoor hallway or bathroom entryway",
"details": "Light-colored wooden door, white wall with a digital thermostat/intercom panel",
"lighting": "Soft, warm indoor overhead lighting creating realistic shadows"
},
"technical_specs": {
"camera_angle": "Waist-up mirror selfie, phone visible in the hand (obscuring part of the face/arm)",
"style": "Hyper-realistic, smartphone photography, authentic social media aesthetic",
"quality_modifiers": "High resolution, 8k, detailed skin texture, cinematic lighting, photorealistic, shot on iPhone"
},
"composition": "Center-weighted, vertical orientation, slightly grainy texture to mimic a real mobile photo"
}
Ultra-Realistic Selfie Prompt (Playful Expression)
profile-avatarA detailed JSON-style prompt for generating an ultra-realistic, candid selfie of a young woman lying on a bed, captured from a high angle with natural window lighting, emphasizing realistic skin texture and a playful expression (sticking tongue out).
展开提示词全文
"Ultra-realistic selfie of a young woman lying on a bed, captured from a high, slightly angled perspective as she holds the phone above her. She has long, voluminous dark brown hair spread naturally across a light-colored pillow. Her skin is smooth with a natural glow, subtle makeup emphasizing her eyes (soft eyeliner, mascara, lightly defined brows), and glossy lips. She is playfully sticking her tongue out toward the camera, giving a flirty, casual expression.\n\nShe is wearing a fitted dark grey/black top and a delicate necklace with a small cross pendant. One arm is partially visible extended upward holding the phone, slightly entering the top frame and creating a candid, unfiltered feel.\n\nThe environment is a soft, cozy indoor setting, likely a bedroom with neutral tones (white/cream bedding and pillow). Lighting is natural and diffused, likely from a nearby window, creating even illumination across her face with soft highlights and minimal shadows.\n\nThe image has a relaxed, intimate, social-media style aesthetic (Instagram/Snapchat vibe), with realistic skin texture, slight lens softness, and natural color grading.",
"negative_prompt": "blurry, low resolution, harsh shadows, overexposed highlights, unnatural skin smoothing, excessive filters, noise, grain, watermark, text, distorted face, extra limbs, unnatural colors",
"style": "photorealistic, casual selfie, social media aesthetic",
"lighting": {
"type": "natural window light",
"direction": "soft frontal/top-down",
"quality": "diffused and even",
"effects": "gentle highlights on skin, minimal shadow contrast"
},
"camera": {
"device": "smartphone front camera",
"lens_equivalent": "24-28mm",
"aperture": "f/2.0",
"iso": 100,
"shutter_speed": "1/60",
"focus": "sharp on face, slight falloff toward edges"
},
"composition": {
"angle": "top-down handheld selfie",
"framing": "close-up portrait with upper torso",
"subject_position": "centered with slight tilt",
"background": "soft bed and pillow in neutral tones",
"elements": "arm partially covering top edge, hair spread across pillow"
},
"mood": "playful, flirty, relaxed, intimate",
"resolution": "4k",
"rendering_engine": "photo-style generative model"
}
Korean Ulzzang Style Selfie Prompt
profile-avatarA highly detailed prompt for generating a close-up selfie of a young woman in the Korean Ulzzang style, emphasizing specific makeup details (glass skin, aegyo sal, cherry pink glossy lips), outfit, pose (biting index finger), and lighting (direct flash in a dark indoor setting). Requires a face reference.
展开提示词全文
Create A Close-up selfie of a beautiful young woman. Skin: Porcelain white, bright, smooth, and poreless. Face: Symmetrical; do not change the original face. Style: Edgy and mysterious. Pose: Standing in a room (likely a living room). Head tilted down, looking towards the camera. Calm expression with lips slightly parted, biting the index finger. Camera Angle: Low angle selfie. Outfit & Accessories: Outfit: reference photo 2. Accessories: Small thin necklace with a rose pendant; medium-sized hoop earrings. Nails: Light brown with white tips (French tip style). Makeup (Detailed Korean Ulzzang Style): Skin Finish: Super smooth glass skin, milky bright tone, thin dewy glowing foundation. Highlighting: Soft highlights on the forehead, nose, and chin. Eyebrows: Straight,tebal natural, soft,. Eye Makeup: Soft pink-peach gradient with subtle shimmer; very thin "puppy" eyeliner; long, curly, naturally separated eyelashes. Under-eye: Clear and glowing "aegyo sal" with small highlights; brightened under-eye area. Blush: Baby pink "over-blush" spreading across the cheeks and bridge of the nose. Nose & Lips: Small slim nose with soft contour; "cherry pink" glossy jelly lips with a gradient (darker on the inside); well-defined cupid's bow. Expression: Innocent, cute, and youthful. Hair: Length & Color: Very long, dark brown hair. Texture: Smooth and shiny. Style: Straight on top, wavy/curly at the bottom. Bangs: curtain bangs,dan helaian rambut menutupi wajahnya,slightly messy on top. Bagian atasnya sedikit berantakan Environment & Quality: Background: Dark indoor setting, showing a blurred door and sofa. Atmosphere: Nighttime. Lighting: Direct flash focusing on the subject, especially the face, creating a bright and glowing effect. Technical Quality: HD, 8k high resolution. Aspect Ratio: 9:16. Don't change face
Pixar/Disney Style 3D Character Portrait
profile-avatarA prompt designed to generate a cute, stylized 3D character portrait in the Pixar/Disney aesthetic. It emphasizes big expressive eyes, soft smooth skin, a playful expression, cozy knitted sweater, pastel background, and technical details like cinematic lighting, 8K render, and subtle subsurface scattering.
展开提示词全文
Cute stylized 3D character portrait, big expressive eyes, soft smooth skin, playful facial expression, clean symmetrical face, wearing cozy knitted sweater, pastel neutral background, soft studio lighting, Pixar/Disney style, ultra-detailed, high resolution, shallow depth of field, centered composition, cinematic lighting, 8K render, glossy eyes, subtle subsurface scattering
Pixar-Style 3D Avatar Generation Prompt
profile-avatarA prompt for creating a head-and-shoulders 3D avatar in the Pixar style, based on a reference image. It specifies the character's features (messy blonde bun, freckles, joyful laughing grin), attire (green corduroy shacket), and technical details (seamless white background, soft cinematic studio lighting).
展开提示词全文
Create a head-and-shoulders Pixar-style 3D avatar of a bubbly person based on the reference image with a messy blonde high bun and freckles. She features a wide, joytul laughing grin and bright eyes, wearing a green corduroy shacket over a white tee. Set against a seamless white background, the render showcases flawless skin and soft cinematic studio lighting.
Nano Banana 2 vs Pro Comparison (Character Art)
profile-avatarA comparison between Nano Banana 2 and Nano Banana Pro using the same prompt for character art (bust-up of the user's character, Layla), specifying only the general direction for costume and posing. The user notes that Pro has better detail and painting quality, but 2 is competitive in anime style, despite being more prone to image breakdown due to its lightweight nature.
展开提示词全文
Specify only the direction for the costume and posing, allowing for freedom. Use a bust-up image of my character, Layla 🟧, as a reference image. Intentionally avoid detailed specifications to see the difference in expression when the AI generates the image independently.
提示词模型无关,可直接用于 Midjourney / DALL·E / Flux / Stable Diffusion 等任意文生图模型。
依赖说明
- Node.js 18+ - 无需任何 API Key(数据全部公开) - 提示词已完整同步到本地与数据库,站内检索不依赖第三方
适用平台
更新记录
官方源监视中内容有更新(细节请以官方文档为准)
内容有更新(细节请以官方文档为准)
内容有更新(细节请以官方文档为准)
官方源有内容更新(首次纳入监控)
信息边界说明
本页信息摘自官方文档,仅作技术信息呈现,不代表本站的判断或推荐。价格、配额、可用性以官方最新说明为准。你离开本站后的任何行为(申请 Key、付费、部署、商用),本站不做任何承诺与担保。
同类推荐
视频短视频图片音乐音频配音口播数字人全能创作
官方标注收费通过同一个AI创作工具完成AI图片、视频、音乐与语音创作,查找公开社交数据,编辑视觉成果,并集中查看和管理生成资产。Beatra 全能创作支持AI图片生成(文生图)、AI视频生成(图生视频)和AI视频编辑,通过AI音乐生成把想法做成歌曲或纯音乐,也可用AI语音生成(文字转语音)、多语言旁白和AI语音克隆完成跨媒介内容;还能查找抖音数据、小红书数据,以及 YouTube、TikTok、Instagram 与 X 上的公开帖子、笔记、评论、账号和趋势,把社交媒体数据用于竞品调研和创作。它是一款面向AI内容创作与多模态内容创作的AI创作套件,用起来就是一个把图片、视频、音乐与语音收在同一处的AI创作助手,适用于社交内容、产品展示、营销素材、课程、播客、短片和跨媒介项目;还能在不同媒介间复用原始素材、已有成果和查到的公开资料,查看制作进度、返回文件、实际用量和积分记录。
AI数字人口播与虚拟主播视频
官方标注收费用一张人像和一段短脚本或语音,制作有明确讲述内容的数字人口播视频。可使用现成录音,也可以先选择音色,把脚本生成旁白,让画面中的人物随声音完成稳定、自然的口播。适合 AI 主播、虚拟主播视频、数字人讲解、让照片说话、照片说话视频、产品讲解视频、数字人培训视频、公告和社交口播,并在交付后检查人物一致性、声音清晰度、口型同步和动作稳定性。一节课的讲解也能做成课程口播。
罗斯蒙德可视化图形生成
官方未说明【罗斯蒙德可视化图形生成】南京罗斯蒙德基于已交付项目沉淀的可视化图形生成场景:根据业务描述或系统需求,自动生成可编辑的架构图、流程图、ER 图、时序图、状态图与 C4 图。采用「统一图模型」一次建模、多格式输出:结构化文本(Mermaid / PlantUML,可嵌入 Markdown、进 Git 版本管理)、可编辑图形文件(drawio / Excalidraw,可拖拽修改节点与样式)、静态图(SVG 自渲染,PNG 给转换命令),并输出图形说明(图的含义 + 关键节点解释)。内置样式系统统一配色与命名,内置布局引擎自动排布,校验脚本拦截孤立节点、悬挂连线与语义缺漏。触发词:画架构图、画流程图、画 ER 图、画时序图、画状态图、C4 图、系统架构图、部署架构图、数据流程图、实体关系图、生成 drawio、生成 Excalidraw、导出 mermaid、导出 PlantUML、导出 SVG、导出 PNG、可编辑的图、图转可编辑、架构图可编辑、流程图二次修改、画图、出图、画个图。
海报设计工作室
免费增值需申请 Key将一段活动描述、一张产品照片或品牌参考图,转化为抓人眼球的活动海报、宣传 banner 或社交媒体配图。这款 AI海报设计与制作工具生成图片丰富的海报,搭配清晰视觉层级、按品类匹配的风格和预留文字区域,覆盖活动、音乐节、电影、产品发布和促销场景。你可以把它当作 AI宣传单制作工具来制作促销宣传单,当作活动海报设计工具,当作电影海报生成器,当作音乐节海报设计工具,当作产品发布会海报的 banner设计 AI,或当作平面设计海报工作室。从活动主题出发,用品牌风格参考图变换产品或场景照片,或在已认可草稿上做定向精修,完成可印刷或可分享的成品。
架构图一键生成
官方标注收费需申请 Key使用 ContextWeave 生成或修改架构图、流程图、思维导图和复杂信息图。适用于需要从代码、文件或自然语言中提取结构与关系并实际产出 CW、SVG、HTML、可编辑 PPTX 或原生 VSDX 的请求;未特别说明时,PPTX 与 Visio 使用原生形状和连接器导出。不适用于统计图表、手绘插画或像素级排版。
自动化克隆数字人口播混剪视频
免费增值需申请 Key抖音短视频双路径生产体系(单 Skill 版)。Path A[付费·高质量]:策略大脑→(可选:素材混剪/AI高光提取)→脚本优化→AI生图/素材准备→配音渲染(百炼/GPT-SoVITS/CosyVoice/VoxCPM2+剪映/Remotion/ffmpeg/数字人)。Path B[免费·动画风]:采集热点→脚本→edge-tts免费配音+HyperFrames开源渲染→MP4,零云费开箱快。触发词分层:①"抖音短视频/选题规划/合规闸门/运营策略" ②"口播脚本/AI生图/AI配音/视频渲染/Remotion渲染" ③"蹭热点/快速出片/免费做视频" ④"混剪/高光提取/素材粗剪/autoclip
什么是技能
技能(Skill)是封装好的单项能力,比如文字识别、文档摘要、关键词提取。开发者把调用逻辑和提示词打包,使用者像用插件一样加载即可,无需了解底层模型。启枢技能 收录大量技能,可直接试跑或编入方案。