← 返回首页
技能

AI Image Prompts · 绘画提示词库

github · 1.0.4 · 下载量 1,092 · 热度分 90.0

出处与授权

作者: YouMind-OpenLab
许可证: MIT 许可证 (原文)
版权声明: Copyright (c) 2026 YouMind-OpenLab

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

作者:YouMind-OpenLab
分类:图文音视频多模态生成
语言:en
版本:1.0.4
认证方式: 无需认证
许可证:MIT
是否付费: 官方标注免费
发布方:YouMind

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。

标签
绘画提示词文生图营销素材

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
适用场景
找某类图片的现成提示词给文章/视频/播客配插画做海报 / 封面 / 社媒图翻译并理解提示词写法

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 702 条

4-Panel Manga Dialogue Prompt

comic-storyboard

A prompt for generating a 4-panel manga titled 'The Truth Beyond the Avatar' with specific dialogue for three characters, generated using Nano Banana Pro (Gemini 3 Pro Image).

展开提示词全文
4-panel manga
Title
The Truth Beyond the Avatar
Man (Center): "Hold on, I got a call from the store. Maybe they're confirming my shift."
Friend A (Left): "Oh, what is it?"
Friend B (Right): "Hope they didn't find out about your fake job, did they?"

Nano Banana Pro 4-Panel Comic Prompt: Bach in Heaven

comic-storyboard

A prompt used with Nano Banana Pro (Gemini 3 Pro Image) to generate a 4-panel comic titled 'Bach in Heaven, Business on Earth' or 'The Light and Shadow of the Urtext'.

展开提示词全文
4-panel comic

Title
"Bach in Heaven, Business on Earth"
​or
"The Light and Shadow of the Critical Edition (Urtext)"

4-Panel Manga: Anonymous and Mask

comic-storyboard

A prompt for Nano Banana Pro (Gemini 3 Pro Image) to generate a 4-panel manga titled 'Anonymous and Mask.' The description specifies a scene inside a moving car where a man in the back seat receives a call from an 'Unknown Number' on his smartphone.

展开提示词全文
4-panel manga
Title

Anonymous and Mask

Description:
 Inside a moving car. The smartphone of the man in the back seat (center of the image) rings. The display shows 'Unknown Number'...

Cinematic Macro Toy-Photography of Mini Beasts

comic-storyboard

A complex, multi-section prompt for generating a cinematic macro toy-photography scene. The concept involves dozens of miniature identical character instances working on a giant open comic book that tells their story. The prompt strictly enforces likeness integrity, specific camera perspective (low-angle macro, 90-100mm lens), shallow depth of field, and hard directional lighting.

展开提示词全文
A cinematic macro toy-photography scene of the uploaded reference character multiplied into dozens of miniature identical instances, each actively working on a giant open comic book that tells the story of the same character.
Likeness Integrity (CRITICAL)

All miniature instances must preserve the exact identity of the reference character:

• identical proportions

• exact eye shape, spacing, stylization

• same silhouette and material behavior

• no added features, no face changes, no stylization drift
The comic book artwork must depict the exact same character design language, translated into printed comic style without altering anatomy or facial structure.
Scene Concept & Action
The giant comic book lies open like a construction site of narrative.
Mini characters are scattered across the pages, each performing a specific task:
• some are drawing panels with tiny tools (pens, brushes, stylus-like tools)

• some are painting color into panels

• some are climbing panel borders like scaffolding

• some are inspecting scenes, pointing, directing

• some are physically interacting with the comic world ( lifting panels, stepping into frames)
The scene must feel like they are building their own story in real-time.
Perspective & Composition (MANDATORY)
Camera & Perspective (Corrected)
Low-angle macro shot, camera positioned close to the comic page
Lens: 90–100mm macro (compression, NOT distortion)
• Perspective compression must preserve consistent object scale

• Foreground character appears larger only due to proximity to lens

• No exaggerated wide-angle distortion

• Spatial depth created through distance falloff + focus, not scaling

• The open comic book stretches diagonally into depth

• Panels form strong leading lines across the frame

• Composition is intentionally off-balance, not centered

• Some characters partially cropped by frame edges
Foreground includes blurred ink tools, paper edges, or panel borders creating depth layering
Depth of Field (CRITICAL)
Extremely shallow depth of field
• Primary focal plane: one or two mini characters in the foreground actively working

• Mid-ground characters slightly softer

• Background characters and panels fall into creamy blur
No full-frame sharpness
Focus must guide the viewer toward the “main worker” character
Lighting Doctrine
Hard directional cinematic light from top-left
• sharp shadows cast across comic panels

• subtle rim light separating characters from page

• ink reflections create small specular highlights

• shadows from panel edges create natural depth
No flat lighting
Environment & Surface Detail
Macro-level realism of the comic book:
• visible paper fibers

• ink texture and slight gloss

• subtle imperfections in print

• panel borders slightly raised or embossed
Tiny tools leave micro mar

Cinematic Portrait Reflection with Ghibli Effect

comic-storyboard

A detailed text prompt for Nano Banana Pro requesting a cinematic portrait of a serious man looking at his reflection in a mirror, emphasizing moody atmosphere, soft natural light, shallow depth of field, and then applying a Ghibli art style effect to the final image.

展开提示词全文
A cinematic portrait of a serious man use image for face reference looking at himself in a mirror, moody atmosphere, soft natural window light coming from the side, dust particles visible in the air, warm wooden interior background, shallow depth of field, sharp focus on the reflection while the back of the man is slightly blurred, emotional and introspective expression, realistic skin texture, short dark hair, light stubble, wearing a casual black jacket over a nike white t-shirt, dramatic lighting, high detail, 50mm lens, photorealistic, film still style. Add Ghibli art effects to this image.

Four-Panel Manga: Judgment of the Ministry of Magic

comic-storyboard

An image generation prompt for Nano Banana Pro (Gemini 3 Pro Image) to create a four-panel manga titled 'Judgment of the Ministry of Magic: Confrontation with Demaon,' referencing Doraemon's 'Great Adventure into the Underworld.'

展开提示词全文
4-panel manga
Title
{argument name="title_en" default="Judgment of the Ministry of Magic: Confrontation with Demaon"}

Demaon
Nobita's Great Adventure into the Underworld

It was supposed to be Demaon
But it didn't seem to match... lol

Four-Panel Manga: Personal Universe

comic-storyboard

An image generation prompt for Nano Banana Pro (Gemini 3 Pro Image) to create a four-panel manga titled 'Personal Universe,' related to the concept of an individual (pillar) and the universe.

展开提示词全文
4-panel manga
Title
{argument name="title_en" default="Personal Universe"}

It's about an individual (pillar) and the universe...

Omitted

Four-Panel Manga: Double Accounting

comic-storyboard

An image generation prompt for Nano Banana Pro (Gemini 3 Pro Image) to create a four-panel manga titled 'Double Accounting,' featuring characters Koto-ge Moto and Katsura.

展开提示词全文
4-panel manga
Title
{argument name="title_en" default="Double Accounting"}

Kotoge Moto
Wig is amazing...

Looks like a different person lol

Static Image Generation from Anime Prompt

comic-storyboard

A prompt used with DeeVid AI's Nano Banana Pro to generate nine static images suitable for a picture-story show (Kamishibai) from an anime prompt titled 'Sharon, the Out-of-Control Girl!'

展开提示词全文
Anime prompt for "{argument name="title" default="Sharon, the Out-of-Control Girl!"}" (『ハチャメチャ暴走娘シャロン!』)

High-Speed Cinematic Cyberpunk Scene Prompt

comic-storyboard

A highly detailed, multi-section prompt designed to turn an uploaded character reference into a photorealistic, high-speed cinematic scene on a rain-slicked cyberpunk street, emphasizing strict identity preservation, specific camera angles, and dynamic pose enforcement, recommended for Nano Banana 2 (Thinking Mode).

展开提示词全文
A cinematic speed-motion shot of the uploaded reference character captured mid-slide on a rain-slicked, neon-drenched cyberpunk street. The focus is equally on the reflection and the character, forming a dual-layer composition. Likeness Preservation (CRITICAL): Maintain exact proportions, eye shape, spacing, stylization, silhouette, and material identity from the reference. Do not humanize. Do not reinterpret anatomy. Do not modify facial structure. Identity must remain perfectly intact and instantly recognizable. 2D to 3D Translation (ONLY if input is 2D artwork): If the uploaded reference is 2D (illustration, anime, flat graphic, cel-shaded), convert it into a fully realized 3D cinematic form while preserving exact identity fidelity. Maintain exact head shape, eye design, spacing, and proportions. Preserve original stylization, do NOT add realism that alters design language. Translate linework into subtle surface geometry or shading transitions. Keep original colors, expressed through physically believable materials. Convert flat shading into controlled cinematic light falloff. The result must feel physically real, not redesigned. Camera & Perspective: Ultra-low angle, camera positioned just centimeters above the wet pavement. Lens: 18–20mm ultra-wide. Slight Dutch tilt (5–10 degrees) to destabilize the frame. Front three-quarter perspective. One limb or body part must be closer to the lens, creating perspective pressure. Composition (Reflection Dominance): The character slices diagonally across the upper half of the frame. The lower half is dominated by a sharp but slightly distorted reflection. Vertical neon signs in the background must mirror into the wet ground, forming strong graphic lines. The composition must feel off-balance and tension-driven, not centered or stable. Pose Enforcement (CRITICAL — OVERRIDE ALL DEFAULT STANCES): The character must NOT be upright, balanced, or casually sliding. The pose must represent loss of balance under high speed. The torso is pitched aggressively forward, close to the ground. One limb forcefully contacts the wet surface, creating splash and friction. The opposite limb extends outward or backward, fully stretched for counterbalance. Legs are asymmetrical: one trailing, one collapsing or stepping incorrectly. Spine forms a diagonal or curved line, never vertical. Head leads the motion slightly ahead of the torso. Weight Distribution: Weight is NOT centered. It collapses toward the contacting limb and forward motion. Ground Interaction (MANDATORY): Water splashes outward from impact points. Friction streaks beneath sliding limb. Surface tension visibly disturbed. Energy Read: The pose must feel like mid-fall recovery, evasive maneuver at high speed, barely controlled momentum. NOT clean skating, surfing, balanced glide, or upright slide. Perspective Amplification: At least one limb must appear larger due to proximity to the lens, enhancing speed and aggression. Motion & Blur: Directional motion blur trailing from the lowest moving limbs. Sharp focal plane on the face or upper torso. Background and trailing limbs fall into controlled blur. Lighting & Color (Cyberpunk Neon): Hard directional neon lighting from off-screen vertical sources. Color palette: deep cyan, magenta, electric yellow. Lighting behavior: sharp highlights on wet surfaces, strong rim light separating silhouette, deep shadow pockets. No flat lighting. Surface Detail (Hyper-Reflective): Wet pavement must feel physically real: mirror-like reflections with ripple distortion, micro scratches and pooling water, subtle rain impact patterns. Character materials slightly darkened from moisture. Atmosphere: Heavy rain and mist in the background. Volumetric light beams cutting through haze. Dense air perspective. Foreground may include subtle water droplets or mist near lens for depth. Emotion: Urgency. Instability. High-stakes motion. The character is either escaping or chasing. Negative Constraints: No centered composition. No upright balanced pose. No clean symmetrical stance. No dry surfaces. No flat reflections. No studio lighting. No evenly sharp full-frame clarity. No static posture. Background Pixel Wall Art (Depth-Controlled): Replace the LED sign with a weathered wall-mounted pixel-art mural or printed panel of the character. The pixel version must preserve silhouette and identity but appear faded, worn, and partially damaged. Printed or painted on a rough wall surface, not glowing like a screen. Subtle grime, scratches, peeling, or moisture distortion. Colors desaturated and affected by environment. Depth Control (CRITICAL): Must sit clearly in the background plane. Slightly obscured by haze, rain, or foreground motion. Edges softened, not crisp. Reduced contrast compared to main subject. Lighting Integration: Does NOT emit light. Only receives ambient neon spill. Partially falls into shadow. Pixel Art Identity Lock (CRITICAL): The pixel-art version must be a direct downscaled translation of the original character, not a reinterpretation. Preserve exact silhouette, head shape, and proportions. Preserve exact eye shape and spacing in simplified pixel form. Use the same facial structure logic — do NOT invent or add features. STRICT PROHIBITIONS: No added nose. No added mouth. No added facial features that do not exist in the original. No stylization drift toward generic pixel characters. If the original character has a minimal or abstract face, the pixel version must remain equally minimal. Foreground Chaos Layer (CRITICAL – DEPTH & IMMERSION): Introduce aggressive near-lens elements that partially obstruct the frame and enhance depth. Near-Lens Elements: Water droplets, splashes, or streaks crossing very close to the camera. Out-of-focus rain particles catching neon light. Small debris or mist fragments passing across the lens plane. These elements must feel physically between the viewer and the character, not part of the background. Depth Behavior (MANDATORY): Foreground elements are heavily out of focus due to proximity. Large soft bokeh shapes or streaks may partially cover edges of the frame. Some elements may briefly occlude parts of the character or reflection. The frame must feel layered: foreground chaos, main subject, background environment. Lens Interaction (IMPORTANT): Subtle water streaks or droplets may appear on the lens surface. Slight distortion or refraction through droplets is allowed. Effects must remain minimal and believable, never covering the focal subject completely. Lighting Interaction: Foreground particles catch neon light (cyan, magenta, yellow). Creates glowing streaks or soft colored flares. Adds dimensional separation between planes. Composition Impact: Frame edges may feel dirty or partially blocked. The image should NOT feel clean or perfectly framed. Slight visual interference is intentional. Emotion: The viewer feels inside the scene, not watching it. ar 16:9

Pop Art comic book illustration prompt for Nano Banana Pro

comic-storyboard

A detailed prompt for Nano Banana Pro to generate a highly stylized, single-panel sequential art illustration in the classic 1960s Pop Art comic book aesthetic (Roy Lichtenstein inspired), featuring a startled woman holding a banana and a man's silhouette in the doorway, complete with Ben-Day dots and a starburst speech bubble.

展开提示词全文
A highly stylized, single-panel sequential art illustration, executed in the classic 1960s Pop Art comic book aesthetic (Roy Lichtenstein inspired). The scene features a startled woman seated on a plush, magenta mid-century modern couch, frozen in an act of surprise. She wears vibrant, high-contrast clothing and is conspicuously holding a single, bright yellow banana. The background shows a door dramatically flung open, revealing a towering, solid black silhouette of a man framed by deep shadows. The entire image is rendered using thick, heavy black ink outlines, exaggerated facial expressions, and pervasive Ben-Day halftone dots. A large, dynamic speech bubble, shaped like a yellow starburst, bursts from the woman's mouth, containing the bold, capitalized text: "{argument name="speech bubble text" default="IT'S NOT WHAT IT LOOKS LIKE!"}" High contrast lighting, vibrant primary color palette, cinematic comic book framing.

4-Panel Manga: The Long Road to 170 Yen Gasoline Price

comic-storyboard

A prompt for Nano Banana Pro (Gemini 3 Pro Image) to generate a 4-panel manga titled 'The Long Road to 170 Yen Gasoline Price,' featuring the comedy duo Reiwa Roman, likely commenting on high gasoline prices.

展开提示词全文
4-panel manga
Title:
The Long Road to 170 Yen Gasoline Price

Reiwa Roman edition

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

claude-codecursorcodexgemini-cliopenclaw

更新记录

官方源监视中
内容更新 2026-10-21

内容有更新(细节请以官方文档为准)

内容更新 2026-10-20

内容有更新(细节请以官方文档为准)

内容更新 2026-10-19

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。

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。

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。

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。