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 1025 条
Photorealistic Fireworks Display Prompt for Dreamina AI
poster-flyerA detailed image generation prompt designed for Dreamina AI to create a photorealistic, long-exposure photograph of fireworks. The fireworks are shaped like a specific subject, clearly visible against a dark night sky above a waterfront city, with colorful reflections on the water and a cinematic composition.
展开提示词全文
Fireworks form the shape of {argument name="subject" default="[SUBJECT]"} in the night sky above a waterfront city. The design is inspired by the colors, proportions, and silhouette of {argument name="subject" default="[SUBJECT]"}, clearly readable and visually striking against the dark sky. Long-exposure, photorealistic fireworks photography with smoke trails, city skyline silhouette, and colorful reflections on the water. Cinematic composition, sharp detail, ultra-high-resolution 4K.
Romantic Cinematic New Year 2026 Beach Night Scene
poster-flyerA prompt for generating a romantic, ultra-realistic New Year 2026 scene on a beach at night, using a strict identity reference. The subject holds a sparkler drawing '2026' against a backdrop of fireworks over the ocean, illuminated by warm sparkler glow and cool moonlight tones.
展开提示词全文
{
"prompt": "Use the uploaded image strictly as the identity reference. Preserve exact face match, facial structure, expressions, skin tone, hairstyle, and body proportions without any alteration. Create a cinematic New Year 2026 beach night scene. The subject stands barefoot or in casual footwear on the beach with a relaxed yet confident pose, wearing a simple stylish outfit such as a {argument name="outfit type" default="white shirt or top with denim or light trousers"} flowing naturally in the ocean breeze. The subject holds a sparkler drawing glowing light trails forming '2026'. Background features a dark ocean shoreline with reflections on wet sand and colorful fireworks lighting up the night sky above the sea. Lighting is cinematic night lighting with warm sparkler glow on the face and cool moonlight tones from the ocean. Shallow depth of field with soft bokeh lights and reflections enhances the dreamy beach mood. Mood is joyful, confident, romantic, and premium New Year celebration. Ultra-realistic photography style, sharp focus, natural skin texture, professional color grading, high dynamic range. Render in ultra high resolution, Instagram and poster ready.",
"negative_prompt": "face mismatch, identity change, extra people, cartoon, anime, illustration, low quality, blur, noise, harsh lighting, extra fingers, extra limbs, distorted anatomy, watermark, logo, text overlay",
"style": "cinematic ultra-realistic night beach photography",
"mood": "romantic, confident, celebratory New Year 2026",
"lighting": "warm sparkler glow with cool moonlight tones",
"background": "beach at night with fireworks over the ocean",
"render_quality": "ultra high resolution, social-media clarity",
"composition": "centered subject, full or mid-body shot",
"reference_strength": 0.9
}
Dark Gothic Cinematic Illustration from Reference
poster-flyerA prompt for transforming an uploaded image's identity into a dark art/gothic cinematic illustration. It requires 100% preservation of facial identity while changing the pose and wardrobe to luxury gothic styles, rendered with bold shadows and high contrast vector art for a dramatic poster feel.
展开提示词全文
{
"prompt": "Use the uploaded image strictly as an identity reference. Preserve 100% exact face match, facial structure, expressions, skin tone, hairstyle, and identity of every person visible in the reference image. Do not add or remove any person. Create a dark art / gothic cinematic pose (for example: standing in shadows, dramatic side profile, intense forward gaze, slow powerful posture) while keeping anatomy natural. Style the subjects in gothic-inspired luxury outfits—long coats, dark fabrics, leather, subtle metallic details—without changing facial identity. Convert the scene into a cinematic illustration blended with clean vector art. Use bold shadows, high contrast, sharp silhouettes, textured gradients, and subtle grain-like artistic shading (not noise) for a dramatic gothic poster feel. Lighting should be low-key cinematic lighting with strong rim light or top light creating depth and mystery. Color palette should be dark and rich—{argument name="color palette" default="black, deep crimson, burgundy, charcoal, muted gold accents"}. Background should be a dark abstract or gothic architectural setting simplified into layered vector forms with fog or smoke elements. Render in ultra high resolution with crisp edges, dramatic contrast, and a polished dark cinematic finish.",
"negative_prompt": "face mismatch, identity change, extra person, missing person, horror gore, anime exaggeration, chibi, wrong facial features, wrong skin tone, bad anatomy, extra fingers, extra limbs, flat lighting, photorealism, low quality, blur, noise, grain, watermark, logo, text",
"style": "dark gothic cinematic illustration + vector art",
"mood": "mysterious, intense, dramatic",
"lighting": "low-key cinematic lighting with rim highlights",
"background": "dark abstract or gothic vector backdrop",
"detail_level": "high dramatic detail",
"render_quality": "ultra high resolution, cinematic poster clarity",
"composition": "same persons, new intense gothic pose",
"reference_strength": 0.9
}
Minimal Cinematic New Year 2026 Portrait
poster-flyerA prompt for generating a minimal, elegant New Year 2026 portrait using a strict identity reference. The subject is shown in a close-up, holding a softly glowing sparkler that subtly suggests '2026', set against a dark, minimal backdrop with soft cinematic lighting.
展开提示词全文
{
"prompt": "Use the uploaded image strictly as the identity reference. Preserve exact face match, facial structure, expressions, skin tone, hairstyle, and body proportions without any alteration. Create a clean minimal New Year 2026 portrait. The subject is captured in a close-up or mid-shot with a calm confident expression, wearing a simple modern outfit in solid neutral colors ({argument name="outfit color" default="white, beige, grey, or black"}). The subject holds a single sparkler softly glowing, forming subtle light trails suggesting '2026' without dominating the frame. Background is dark or softly blurred with minimal bokeh lights, no fireworks. Lighting is soft cinematic with gentle highlights on the face and warm sparkler glow adding depth. Mood is elegant, confident, modern, and premium New Year celebration. Ultra-realistic photography style, natural skin texture, professional color grading, sharp focus, high dynamic range. Render in ultra high resolution, perfect for profile photo, banner, and minimal social media post.",
"negative_prompt": "face mismatch, identity change, extra people, cartoon, anime, illustration, low quality, blur, noise, harsh shadows, extra fingers, extra limbs, distorted anatomy, fireworks, confetti clutter, watermark, logo, text overlay",
"style": "minimal cinematic realistic photography",
"mood": "clean, confident, elegant New Year 2026",
"lighting": "soft cinematic lighting with subtle sparkler glow",
"background": "dark minimal backdrop with soft bokeh",
"render_quality": "ultra high resolution, clean social media quality",
"composition": "close-up or mid-shot, centered face focus",
"reference_strength": 0.9
}
Figure Product Introduction Layout Prompt (Multilingual)
poster-flyerA Japanese tweet discussing the 'Figure Conversion + Product Introduction Layout Prompt' (Prompt 13 from the festival list) and its multilingual capability (Japanese, English, German, French). The user notes that the prompt allows for the insertion of catchphrases.
展开提示词全文
Let's play with Nano Banana Pro [Figure Product Introduction Style Layout Prompt Multilingual Version]
Magazine Page Generation for Travel Guide
poster-flyerA simple prompt used to test the ability of Nano Banana Pro to generate a magazine page layout containing information about a specific tourist spot, in this case, the 'Ijinkan' (Foreign Residences) in Kobe.
展开提示词全文
Generate the "{argument name="tourist spot" default="Ijinkan"} page" of a magazine filled with information for guiding tourists in Kobe
Drone Art Character Projection over Night Sea
poster-flyerA prompt for generating a realistic photo of a character's shape being formed in the night sky over the sea by thousands of glowing drones, creating a spectacular, fantastical illumination event captured with a wide-angle lens. It uses a reference image to define the character's features.
展开提示词全文
[Features of the reference image: Input Image] A realistic photo where the shape of the character with the above features is formed in the night sky by countless glowing drones floating over the sea. The location is the night sky above the sea. Thousands of drone LED lights are arranged like pointillism in the dark sky, drawing the character in neon colors in the night sky. The light from the drones is reflected and shimmering on the dark sea surface below. A magnificent and fantastical illumination event landscape, photographed with a wide-angle lens.
Futuristic New Year Magazine Cover Design Prompt
poster-flyerA highly structured prompt for Nano Banana Pro to generate a high-end, tech-visual magazine cover with a New Year's motif. It specifies a complex layered composition, including a white background with subtle gold leaf, a dynamic anime mascot in a kimono-suit hybrid, and precise typography for the masthead and feature titles, all adhering to a modern Japanese graphic design aesthetic.
展开提示词全文
metadata:
title: "Nano Banana Pro - New Year Special 2026 Edition"
model_context: "High-End Japanese Modern Magazine Cover - Tech-Visual Aesthetic"
image_structure:
composition_type: "Magazine Cover Layout with New Year Motif"
visual_flow: "Top Masthead -> Dynamic Character in Kimono-Style Suit -> New Year Slogan -> Technical Details"
layers:
layer_background:
shape: "Full screen background."
color: "Pure White (#FFFFFF)"
texture: "Subtle gold leaf texture or geometric 'Asanoha' (hemp leaf) patterns in faint light grey. A large, stylized red circle (Rising Sun) positioned offset to the left."
lighting: "Sharp studio lighting with golden rim light."
layer_masthead_top:
location: "Top section."
elements:
- type: "Small Top Text"
text: "NEW ERA | CREATIVE REVOLUTION | {argument name="year" default="2026 AD"}"
style: "Spaced-out, gold-colored elegant sans-serif text."
- type: "Main Brand Logo"
text: "NANO BANANA PRO"
style: "Massive, bold, matte black text. The character's hair or katana tip overlaps the letters."
- type: "Special Edition Subtitle"
text: "謹賀新年 - THE ART OF PROMPTING"
style: "Vertical Japanese calligraphy style on the right side, sleek and modern."
layer_character_main:
location: "Center, dynamic silhouette."
content: "Cool, futuristic Japanese Anime Girl (Nano Banana Pro mascot)."
style: "Ultra-detailed 2D anime art with cinematic shading."
details:
subject: "A creative warrior ready for the new year."
appearance:
hair: "Platinum white hair with vibrant banana-yellow highlights."
clothing: "A fusion of a traditional Kimono and a tactical pilot suit. White base with red accents and gold embroidery. Cybernetic 'Obi' (belt) with glowing neon yellow lights."
accessories: "Stylized katana made of translucent yellow glass (energy blade)."
pose: "A sharp 'Iai' pose or a mid-air dash, looking directly at the viewer with confidence. The kimono sleeves create a wide, dramatic flow across the canvas."
layer_foreground_text:
elements:
- type: "Main Feature Title"
text: "N E W S T A N D A R D"
location: "Lower third, overlapping character's legs."
style: "Large, bold black typography. The text has a 'glitch' effect with gold particles."
- type: "Teaser Text Block (Left)"
location: "Middle Left."
text: "New Year, creativity accelerates.\nTo the pinnacle of AI design.\nThe unexplored realm of expression,\nin your hands."
style: "Clean, vertical Japanese text, dark grey for high legibility."
- type: "Footer Info (Bottom Right)"
location: "Bottom Right corner."
content:
header: "CONTENTS"
list:
- "2026 Innovation in Production Flow"
- "The Secret of the Golden Ratio Prompt"
- "AI x Japanese Design Aesthetics"
issue_number:
text: "VOL. 026 | NEW YEAR SPECIAL"
style: "Black rectangle with white/gold text, looking like a premium seal."
style_and_tone:
art_style: "Modern Japanese Graphic Design"
color_palette:
primary: "White, Deep Red, Matte Black"
secondary: "Electric Yellow (Banana), Metallic Gold"
quality_tags:
- "Masterpiece"
- "New Year visual"
- "Sharp typography"
- "Modern Japanese aesthetic"
- "Ray tracing reflections"
negative_prompt:
- "tradit
New Year's Card Illustration with Seven Lucky Gods Costume Swap
poster-flyerA detailed prompt for generating a New Year's card illustration in a cute, mini-character (SD character) style. It requires the AI to take up to seven input characters, randomly assign them the roles and traditional costumes of the Seven Lucky Gods, and fill in any missing characters to complete the set of seven. The final image must include specific calligraphy text, a treasure ship, Mount Fuji, and the first sunrise.
展开提示词全文
Please set the image size to "{argument name="image size" default="Horizontal postcard size (aspect ratio 3:2)"}".
Create a New Year's card illustration featuring "7 mini characters (chibi characters)" riding a treasure ship.
[Character Specification (Clothing must be changed)]
(*Enter character features here)
- Brown hair, short twin tails, wearing a black cloak
- Long black hair, girl in a sailor uniform
- Shoulder-length hair covering one eye, black hair, boy with a hood
[Generation Rules]
1. **Style**: Draw everyone in a cute "mini character (SD character)" style, 2 to 3 heads tall. They should have an adorable form with a large head and small body.
2. **Random Assignment**: Randomly assign the characters the roles of the Seven Lucky Gods (Ebisu, Daikokuten, Benzaiten, Bishamonten, Hotei, Jurōjin, Fukurokuju), ensuring **no duplication and one role per person**. The image must contain exactly one of each of the 7 distinct Lucky Gods' costumes.
3. **COSTUME MANDATE**: All characters (including the User Characters) MUST wear "Traditional Seven Lucky Gods Costumes" (Kimono, Robes, Armor).
4. **PROHIBITED**: NO modern clothes. NO hoodies. NO t-shirts. NO shorts. NO suits. NO glasses (unless specified).
5. **AUTO-FILL**: If the user list is fewer than 7, fill the remaining spots with the original Gods. Ensure the filler Gods have distinct faces (e.g., old bearded men, fat happy man) to avoid everyone looking like a young anime character.
[Text Information]
Clearly draw the following text once each in a **masterful brushstroke style** within the illustration:
1. "Gashou" (Happy New Year)
2. "2026"
[Background/Atmosphere]
The treasure ship is overflowing with gold and treasures. The background features Mount Fuji and the first sunrise. The overall illustration should be bright, pop, and full of New Year's celebratory mood.
Generating a Magazine Cover for Edogawa Ward Tourism
poster-flyerA user tested a prompt asking Nano Banana Pro to generate a magazine cover filled with information to guide tourism in Edogawa Ward, comparing the output with ChatGPT. The prompt is a text-to-image instruction for a specific editorial design.
展开提示词全文
Generate a magazine cover filled with information to guide tourism in Edogawa Ward.
Minimalist Editorial Illustration Prompt
poster-flyerA simple prompt for generating minimalist, editorial-style illustrations. It focuses on flat colors, strong negative space, and clean lines, suitable for a magazine cover aesthetic.
展开提示词全文
A minimalist illustration showing a small {argument name="person description" default="PERSON"} standing beneath a massive {argument name="symbolic shape" default="SYMBOLIC SHAPE"}. Flat colors, strong negative space, clean lines, editorial magazine cover aesthetic.
Superhero Comic Book Pop Art Diorama Prompt
poster-flyerA detailed prompt for generating a pop art-style image of a superhero (like Iron Man) bursting out of a graphic novel page in 3D, set on a messy desk environment. It uses comic book aesthetics like cel-shading and halftone patterns.
展开提示词全文
do this for {argument name="superhero name" default="Iron Man"} <instructions>Subject: An open graphic novel. A superhero character is doing his signature move through a page in 3D, with the right sound effect bubbles floating in the air.
Style: Pop art mixed with character appropriate style, comic book cel-shading, vibrant primary colors, halftone patterns.
Environment: A messy [audience appropriate] desk with action figures and soda cans related to the this super hero's universe.
Lighting: Bright, flat, high-key lighting typical of comic book coloring mixed with character appropriate colors.
Composition: Dynamic, foreshortened perspective emphasizing the signature move coming toward the viewer.</instructions>
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。