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 224 条
Majestic Elephant Cloud Photography
youtube-thumbnailA surreal prompt creating hyper-realistic photography of volumetric clouds formed into a detailed elephant over a savanna.
展开提示词全文
Hyper-realistic photography of massive volumetric clouds forming the shape of a majestic {argument name="animal" default="elephant"}, highly detailed and anatomically accurate cloud structure with enormous tusks and a raised trunk, soft yet dense cumulus clouds with realistic depth and texture, dramatic golden-hour lighting, sunlight rays piercing through the clouds, subtle shadows and highlights enhancing form.
Foreground: a vast {argument name="landscape" default="African savanna meadow"} with realistic dry grass, atmospheric perspective, and cinematic depth of field. Multiple {argument name="wildlife" default="zebras"} grazing naturally across the field. Mid-air: flock of birds circling around the giant elephant cloud, adding scale and motion.
Environment: cinematic sky with warm orange and blue gradient tones, slight haze, natural color grading, ultra-realistic lighting, global illumination, volumetric fog, physically accurate shadows.
Style: DSLR photography, 85mm lens, f/2.8 aperture, shallow depth of field, HDR, 8K resolution, razor-sharp focus, intricate details, photorealism, no CGI look.
Mood: majestic, calm, surreal yet believable, wildlife documentary realism.
Anime Railway Station Transformation
youtube-thumbnailTransforms a photo or sketch of a wooden railway station into an anime-style illustration with descriptive prompts.
展开提示词全文
Please create it in an anime style. {argument name="scene description" default="A wooden station building with a train stopped in the distance."}
Mr Bean Mini Cooper Chaos
youtube-thumbnailA hyper-realistic wide-angle shot of Mr. Bean driving a vintage Mini Cooper maniacally with animated characters integrated into a chaotic scene.
展开提示词全文
Hyper-realistic, chaotic wide-angle fish-eye dashboard shot of a speeding {argument name="car" default="vintage Mini Cooper"}.
Characters & Identity:
{argument name="driver" default="Mr. Bean (Rowan Atkinson)"} driving maniacally, steering with knees, holding a cup of tea.
Passenger: photorealistic human, face and hair 100% identical to the uploaded reference, wearing a {argument name="outfit" default="modern oversized printed baggy t-shirt"}, enjoying the thrill of speed.
Back seat: 3D animated Chucky (evil doll) and Alien Xenomorph, physically squished together, faces pressed against the side window in humorous distortion.
Outside windshield: 3D animated Green Mean Gremlin splattered against glass, furious expression.
Through rear window: police car in pursuit.
Scene & Environment:
atermark, text Harsh daylight, strong natural shadows
Motion blur on passing scenery for extreme sense of speed
Interior of Mini Cooper rendered with photorealistic detail (dashboard, steering wheel, seats, reflections)
Camera & Composition:
Wide-angle fish-eye perspective, exaggerating speed and chaos
Cinematic framing, dynamic, immersive POV
High-fidelity textures, realistic lighting, reflections, and depth
Mood & Realism: Hyper-realistic photorealism for all human and Mini Cooper elements
3D animated characters integrated seamlessly into scene
Vibrant, chaotic, fast-paced, cinematic, humorous
Cinematic Moonlight Rooftop Portrait Collage
youtube-thumbnailA sophisticated multi-frame prompt that creates an emotional, film-like collage of a person on an urban rooftop under a massive full moon, focusing on mood and atmospheric textures.
展开提示词全文
Use the uploaded face reference image as the exact identity reference. Maintain perfect facial consistency across all frames — same facial structure, skin tone, hairstyle, eye shape, beard texture, nose, lips, and realistic proportions. Do not change identity. Create an ultra-cinematic multi-frame night photography collage of a {argument name="subject" default="young man"} standing on a {argument name="location" default="rooftop"} under a {argument name="celestial body" default="massive glowing full moon"}. Mood should feel emotional, dreamy, introspective, and film-like. Composition Layout (6-frame cinematic grid): 1.Wide rooftop shot with giant moon dominating the cloudy night sky. 2.Side profile portrait looking upward toward the moon. 3.Extreme close-up of eye reflecting moonlight and clouds. 4.Cinematic silhouette frame from behind with skyline lights. 5.Hand close-up resting on rooftop railing with shallow depth of field. 6.Atmospheric sky frame filled with stars, clouds, moon glow, and haze. Environment: Urban rooftop at night, Distant city skyline with realistic bokeh lights, Dramatic clouds surrounding bright moon, Slight wind movement in clothes and hair, Moist rooftop texture with reflective highlights, Atmospheric fog and cinematic haze. Lighting: Soft blue moonlight as key light, Subtle warm city lights from below, Realistic night shadows, Volumetric moon glow, Rim lighting around silhouette, High dynamic range cinematic contrast. Camera & Photography Style: Shot on Sony A7R IV / ARRI Alexa style, 85mm portrait lens + 35mm cinematic wide lens, Ultra photorealistic, Natural skin texture, Shallow depth of field, Sharp eye detail, Film grain, HDR night photography, Moody color grading, Premium cinematic composition, Hyper detailed, Unreal Engine realism, Netflix-style visual atmosphere. Color Palette: Deep midnight blue, silver moonlight, soft black shadows, cool cyan tones, warm golden city bokeh highlights. Quality Tags: masterpiece, best quality, ultra detailed, 8k, photorealistic, cinematic lighting, volumetric fog, realistic skin texture, atmospheric perspective, dramatic composition, high contrast, emotional storytelling, movie still aesthetic, premium cinematic collage, highly detailed clouds, realistic reflections. Negative Prompt: low quality, blurry face, bad anatomy, distorted eyes, extra fingers, oversaturated colors, cartoon, CGI look, duplicate frames, low detail skin, noisy image, warped hands, fake lighting, washed colors, deformed face, unrealistic moon proportions, text, watermark, logo, cropped face
Street Glamour Cake Knife Portrait
youtube-thumbnailA cinematic high-angle portrait of a woman in Balenciaga-style clothing sitting on a carpet with a cake, licking a blade for a relaxed street glamour vibe.
展开提示词全文
Do not change the facial features. A cinematic close-up portrait of a young woman. The photo is taken from a slightly high angle. She is sitting on the floor, which is covered with a gray carpet. She is facing the camera; one leg is bent, the other is straight and extended. Between her legs, there is a flat glass plate with a partially eaten cake on it. One hand is next to the cake and is smeared with it; with the other hand, she is holding a knife and licking the cake off the blade. Her gaze is slightly turned to the side. The photo is taken with flash. Film grain is added. The overall vibe of the frame is relaxed, with a street glamour feel. She is wearing {argument name="clothing style" default="Balenciaga-style clothing"}. A realistic, detailed photograph.
Cinematic Vigilante Graphic Poster
youtube-thumbnailA high-contrast cinematic poster prompt featuring a masked vigilante in a tactical suit with bold typography and a gritty noir aesthetic.
展开提示词全文
A cinematic, vertical graphic poster of a masked vigilante in a tactical carbon-fiber suit. The character is positioned in the foreground with a gritty, matte black texture. The background features massive, bold stencil-style typography in a deep crimson red with heavy metallic scratching and weathered textures. The lighting is low-key and dramatic, with subtle rim lighting on the suit to define its armor plates. High-contrast noir aesthetic, moody atmosphere, sharp details, 8k resolution, Extreme macro close-up of a monstrous green-skinned face, focused on the nose, furrowed brow, and intense glowing green eyes. The skin is hyper-detailed with deep wrinkles, sweat, and a rough, leathery texture. Overlaid vertically across the center is the word "HULK" in large, bold, weathered white sans-serif typography with a metallic, scratched finish. The color palette is dominated by deep emerald greens and shadows, featuring explosive green energy particles and debris flying across the frame. Cinematic lighting, gritty atmosphere, 8k resolution, hyper-realistic, aggressive and powerful mood.
Basketball Arena Broadcast Portrait
youtube-thumbnailA complex structured prompt in JSON format to generate a broadcast-style image of a woman at a basketball game, complete with UI overlays and arena lighting.
展开提示词全文
usw image1 as the main identity anchor
{"meta":{"image_quality":"High","image_type":"Screenshot","resolution_estimation":"1920x1080","file_characteristics":{"compression_artifacts":"Low","noise_level":"Low","lens_type_estimation":"Medium telephoto"}},"global_context":{"scene_description":"A medium shot of a {argument name="subject" default="young woman with long dark hair"} sitting in a crowded arena, likely a basketball game, with an {argument name="overlay type" default="ESPN broadcast overlay"} at the bottom and top right.","environment_type":"Indoor arena","time_of_day":"Unknown","weather_atmosphere":"Crowded, indoor lighting","lighting":{"source":"Artificial arena lighting","direction":"Frontal, slightly overhead","quality":"Soft, diffused","color_temperature":"Neutral"},"color_palette":{"dominant_hex_estimates":["#1a1a1a","#333333","#ffffff"],"accent_colors":["#fdb927","#007a33"],"contrast_level":"Medium"}},"composition":{"camera_angle":"Eye-level","framing":"Medium shot","depth_of_field":"Shallow","focal_point":"The woman's face","symmetry_type":"None","rule_of_thirds_alignment":"Woman's face is positioned in the upper right quadrant"},"objects":[{"id":"obj_001","label":"Woman","category":"Person","location":{"relative_position":"Center","bounding_box_percentage":{"x":0.25,"y":0.05,"width":0.5,"height":0.9}},"dimensions_relative":"Large","distance_from_camera":"Near","pose_orientation":"Facing slightly right, head turned towards the left","material":"Skin, hair, fabric","surface_properties":{"texture":"Smooth skin, wavy hair","reflectivity":"Low","micro_details":"Natural skin texture, subtle makeup","wear_state":"N/A"},"color_details":{"base_color_hex":"#d2b48c","secondary_colors":["#000000"],"gradient_or_pattern":"None"},"interaction_with_light":{"shadow_casting":"Soft shadows under chin and nose","highlight_zones":"Cheekbones, nose bridge, lips","translucency":"None"},"text_content":null,"relationships":[]},{"id":"obj_002","label":"ESPN Scoreboard Overlay","category":"UI elements","location":{"relative_position":"Bottom center","bounding_box_percentage":{"x":0.25,"y":0.85,"width":0.5,"height":0.15}},"dimensions_relative":"Medium","distance_from_camera":"Overlay","pose_orientation":"Flat","material":"Digital graphic","surface_properties":{"texture":"Smooth","reflectivity":"None","micro_details":"Sharp typography","wear_state":"N/A"},"color_details":{"base_color_hex":"#000000","secondary_colors":["#fdb927","#007a33","#ffffff"],"gradient_or_pattern":"Solid color blocks"},"interaction_with_light":{"shadow_casting":"None","highlight_zones":"None","translucency":"Slightly opaque"},"text_content":{"raw_text":"{argument name="scoreboard text" default="LAL 51 BOS 43 2ND 3:31 24 NBA WEDNESDAY"}","font_style":"Sans-serif","font_weight":"Bold","text_case":"Uppercase","alignment":"Center","color_hex":"#ffffff"},"relationships":[]}],"background_details":{"texture":"Blurred crowd","patterns":"None","lighting_behavior":"Out of focus","additional_elements":["Person holding a plastic cup on the left"]}}
NBA Broadcast Stadium Screenshot
youtube-thumbnailA complex structured prompt designed to simulate an ESPN broadcast screenshot of a woman in a basketball arena crowd.
展开提示词全文
{"meta":{"image_quality":"High","image_type":"Screenshot","resolution_estimation":"1920x1080","file_characteristics":{"compression_artifacts":"Low","noise_level":"Low","lens_type_estimation":"Medium telephoto"}},"global_context":{"scene_description":"A medium shot of a young woman with long dark hair sitting in a crowded arena, likely a basketball game, with an ESPN broadcast overlay at the bottom and top right.","environment_type":"Indoor arena","time_of_day":"Unknown","weather_atmosphere":"Crowded, indoor lighting","lighting":{"source":"Artificial arena lighting","direction":"Frontal, slightly overhead","quality":"Soft, diffused","color_temperature":"Neutral"},"color_palette":{"dominant_hex_estimates":["#1a1a1a","#333333","#ffffff"],"accent_colors":["#fdb927","#007a33"],"contrast_level":"Medium"}},"composition":{"camera_angle":"Eye-level","framing":"Medium shot","depth_of_field":"Shallow","focal_point":"The woman's face","symmetry_type":"None","rule_of_thirds_alignment":"Woman's face is positioned in the upper right quadrant"},"objects":[{"id":"obj_001","label":"Woman","category":"Person","location":{"relative_position":"Center","bounding_box_percentage":{"x":0.25,"y":0.05,"width":0.5,"height":0.9}},"dimensions_relative":"Large","distance_from_camera":"Near","pose_orientation":"Facing slightly right, head turned towards the left","material":"Skin, hair, fabric","surface_properties":{"texture":"Smooth skin, wavy hair","reflectivity":"Low","micro_details":"Natural skin texture, subtle makeup","wear_state":"N/A"},"color_details":{"base_color_hex":"#d2b48c","secondary_colors":["#000000"],"gradient_or_pattern":"None"},"interaction_with_light":{"shadow_casting":"Soft shadows under chin and nose","highlight_zones":"Cheekbones, nose bridge, lips","translucency":"None"},"text_content":null,"relationships":[]},{"id":"obj_002","label":"ESPN Scoreboard Overlay","category":"UI elements","location":{"relative_position":"Bottom center","bounding_box_percentage":{"x":0.25,"y":0.85,"width":0.5,"height":0.15}},"dimensions_relative":"Medium","distance_from_camera":"Overlay","pose_orientation":"Flat","material":"Digital graphic","surface_properties":{"texture":"Smooth","reflectivity":"None","micro_details":"Sharp typography","wear_state":"N/A"},"color_details":{"base_color_hex":"#000000","secondary_colors":["#fdb927","#007a33","#ffffff"],"gradient_or_pattern":"Solid color blocks"},"interaction_with_light":{"shadow_casting":"None","highlight_zones":"None","translucency":"Slightly opaque"},"text_content":{"raw_text":"{argument name="scoreboard text" default="LAL 51 BOS 43 2ND 3:31 24 NBA WEDNESDAY"}","font_style":"Sans-serif","font_weight":"Bold","text_case":"Uppercase","alignment":"Center","color_hex":"#ffffff"},"relationships":[]}],"background_details":{"texture":"Blurred crowd","patterns":"None","lighting_behavior":"Out of focus","additional_elements":["Person holding a plastic cup on the left","Person on the right"]}}
Magical Book Visualization Grid
youtube-thumbnailA creative weighted prompt formula for generating a 2x2 grid of fantasy dioramas inside open books.
展开提示词全文
2x2 grid, do this for 4 famous books render_target = ( {argument name="main object" default="giant_sliced_open_hardcover_fantasy_book"} * 2.0 ) + ( {argument name="inner scene" default="inferred_magical_library_diorama_inside"} * 1.8 ) + ( miniature_main_characters_fighting_bad_guys * 1.5 ) + ( floating_old_english_map_overlay_top_right * 2.0 ) + ( tiny_ink_bottles_and_quills_exterior * 1.0 ) + ( {argument name="background" default="clean_parchment_paper_studio_background"} * 1.2 ) - ( e_reader / 2.0 ) - ( modern_clothing / 2.0 ) - ( messy_desk )
Song-to-Image Visualization
youtube-thumbnailA prompt that uses a specific song title and artist to generate an atmospheric image representative of the music.
展开提示词全文
{argument name="song title" default="Istanbul Mambo"} ({argument name="artist" default="Eiichi Ohtaki"})
Cinematic Podcast Studio Portrait
youtube-thumbnailA prompt for a professional podcast host in a modern studio setup, featuring warm accent lighting and realistic equipment details.
展开提示词全文
A professional {argument name="host description" default="podcast host sitting at a wooden desk in a modern studio setup"}, facing the camera with a confident and friendly expression. He is wearing a {argument name="clothing" default="fitted black shirt"} and sitting in a comfortable chair. A high-quality podcast microphone on a boom arm is positioned close to his mouth. On the desk are a notebook with handwritten notes, a coffee mug, and an audio interface with connected cables. The background features {argument name="background lighting" default="stylish indoor plants and warm red-orange accent lighting"}, creating a clean, cinematic studio atmosphere. Soft key lighting on the face, subtle rim lighting, shallow depth of field, ultra-realistic details, 50mm lens, f/1.8, professional color grading, high contrast, 8K, YouTube podcast aesthetic.
Japanese TV Program Style Image Generation
youtube-thumbnailA prompt template for Nano Banana Pro to generate images in the style of a Japanese TV program, allowing the AI to determine the overall atmosphere, including captions and interview content, based on a brief scenario.
展开提示词全文
Generate an image in the style of a Japanese TV program.
Let the AI consider the overall atmosphere of the image, including captions and interview content/answers associated with the following details:
{argument name="theme" default="NISA poverty. Morning news program. A 20-year-old female idol being interviewed at a woman's home."}
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
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。
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。
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。
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。
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。
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。