← 返回首页
技能

Ai Image Prompts

tencent_skillhub · 1.0.9 · 下载量 1,827 · 热度分 55.6

产品介绍

从包含10,000+真实图像生成提示的精选库中推荐提示。兼容任何AI图像模型,包括Nano Banana Pro、Nano Banana 2、Seedream等。

资源信息

作者:dophinl
分类:图文音视频多模态生成
版本:1.0.9
认证方式: 无需认证
是否付费: 官方未说明

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
图片生成

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 725 条

GTA 6 In-Game Footage Test Prompt

game-asset

A test prompt used to compare Nano Banana 2 and New GPT-Image, requesting highly realistic, detailed GTA 6 in-game footage, shot from a stationary 4k monitor, with a character walking on the beach with a dog.

展开提示词全文
GTA 6 in-game footage, very detailed, very realistic. Close-up shot taken from a stationary 4k monitor. (There's a slight blurriness in the image, as it feels like it was taken handheld). A wide, bright environment. Realistic details. The character is walking on the beach with /  :dog.

Nano Banana 2 image editing prompt

game-asset

A prompt used with Nano Banana 2 to edit an existing screenshot, instructing the AI to make the modern graphics look like a 90s boomer shooter game.

展开提示词全文
Take this screenshot and edit it. The graphics are too modern. please make this look like a 90s boomer shooter.

TCG style pole-mounted transformer

game-asset

A prompt used with Nano Banana Pro (Google Gemini 3.1 image generator) to create an image of a pole-mounted transformer in the style of a trading card game card.

展开提示词全文
a pole-mounted transformer in the style of a trading card game's card

Action Pose Sheet for Cat

game-asset

A prompt for generating an action pose sheet for the stylized 3D animated lanky grey housecat, including specific poses like mid-pounce and skidding stop, emphasizing exaggerated slapstick silhouettes.

展开提示词全文
Character turnaround sheet of the same lanky grey housecat with oversized paws, cream muzzle, amber eyes, expressive black brows, tall triangular ears, slim elastic body, red collar. Show full body front view, side view, back view, and three-quarter view in one clean composition. Keep proportions identical in every angle. Tail fully visible. Neutral standing pose, arms relaxed, readable silhouette. premium stylized 3D family-film animation, rounded appealing shapes, soft groomed fur, clean neutral studio background, no props, no text, no watermark. Character turnaround sheet of the same tiny warm-brown mouse with round ears, quick black eyes, pale belly, tiny pink nose, slim tail, and tiny blue scarf. Show full body front view, side view, back view, and three-quarter view in one clean composition. Keep scarf shape, tail length, ear size, and body proportions identical in every angle. Neutral standing pose, readable silhouette. premium stylized 3D family-film animation, clean neutral studio background, no props, no text, no watermark. Action pose sheet of the same tiny warm-brown mouse with round ears, quick black eyes, pale belly, tiny pink nose, slim tail, and tiny blue scarf. Show full-body poses in one image. Exaggerated but clean slapstick silhouette, consistent anatomy and scarf shape. premium stylized 3D family-film animation, neutral studio background, no text, no watermark. Action pose sheet of the same lanky grey housecat with oversized paws, cream muzzle, amber eyes, expressive black brows, tall triangular ears, slim elastic body, red collar. Show full-body poses in one image: sneaking low, mid-pounce, skidding stop, hanging from one paw, pancake-splat pose, triumphant pose. Exaggerated but clean slapstick silhouette, consistent anatomy, same proportions in every pose. premium stylized 3D family-film animation, neutral studio background, no text, no watermark

Hyper-Realistic 3D Chess Set for Rivalry Themes

game-asset

A highly structured Nano Banana 2 prompt template for generating a hyper-realistic, isometric 3D render of a chess set based on a rivalry theme, such as Terminator vs. Bleach. It defines strict rules for board geometry, material morphing, army placement (Foreground vs. Background), piece hierarchy, and specific lighting/color contrast for each faction.

展开提示词全文
Input Variable:   [{argument name="rivalry theme" default="INSERT RIVALRY HERE"}] (e.g., Marvel vs. DC)

  System Instruction:  
Generate a hyper-realistic, isometric 3D render of a Chess Set.   CRITICAL:   You must adhere to strict spatial rules regarding Army Placement and Terrain.

  1. Container (The Battle-Scarred Board):  
      Geometry:   A strict 8x8 Euclidean Grid.
      The Divide (The Diagonal Morph):   The terrain morphs diagonally from the   Bottom-Left Corner   to the   Top-Right Corner  .
         The Foreground Territory (Bottom-Left):  [{argument name="material A" default="Material A"}] (e.g., Marvel Tech/Stark Tower Glass).
         The Background Territory (Top-Right):  [{argument name="material B" default="Material B"}] (e.g., Gotham Concrete/Gargoyle Stone).
      The Seamless Transition:   The squares are distinct, but the material morphs diagonally across the board.

  2. The Roster (Strict Foreground vs. Background):  
  DO NOT   mix the armies.   DO NOT   split them Left/Right.
      Army A (The Protagonist - Foreground):  
         Placement:  Occupies the   Bottom 2 Rows   closest to the camera.
         The Pawns:  8 Identical Units (Row 2).
         The Elites:  King, Queen, Rooks, etc. (Row 1).
         Identity:  MUST be strictly Faction A (e.g., Only Avengers/Iron Man style).

      Army B (The Antagonist - Background):  
         Placement:  Occupies the   Top 2 Rows   furthest from the camera.
         The Pawns:  8 Identical Units (Row 7).
         The Elites:  King, Queen, Rooks, etc. (Row 8).
         Identity:  MUST be strictly Faction B (e.g., Only Justice League/Batman style).

  3. The Piece Hierarchy :  
      The King:   A Single, Tall, Unique Artifact per side (Center of the back row).
      The Queen:   A Single, Sharp/Elegant Artifact per side (Next to King).
      The Pawns:   Uniform, small "minion" pieces.

  4. Visual :  
      Camera:     High-Angle Isometric.   Looking down at the board diagonally so the "Marvel Corner" is closest to the viewer and "DC Corner" is furthest away.
      Contrast:   The Foreground pieces (Marvel) should be illuminated by [{argument name="color A" default="Color A"}] (e.g., Cyan Arc Reactor glow). The Background pieces (DC) should be illuminated by [Color B] (e.g., Amber Streetlight glow).

  Output:   A legal chess setup with armies facing off, 8k resolution, clean composition.

Isometric Luxury Product Showcase of a Vinyl Collectible Figure

game-asset

A highly specific prompt for generating an isometric luxury product showcase of a person (from a reference photo) reimagined as a premium vinyl collectible figure set. It details the figure's chibi proportions, the window box packaging, curated accessories, and professional commercial toy photography lighting specifications.

展开提示词全文
Isometric luxury product showcase of the person from the reference photo reimagined as a premium vinyl collectible figure set, stylized chibi proportions with an oversized head and refined facial likeness, displayed in an elegant collector’s window box beside the actual figure posed outside the packaging, with curated accessories reflecting the subject’s identity arranged neatly inside the box tray. Packaging with a clean title area for {argument name="name" default="[Name]"}. High-end commercial toy photography, shot on full frame with a 50mm lens at f/8, ISO 100, 1/160s. Controlled studio lighting with a large softbox key at 45° camera-left, subtle fill bounce, stripbox edge light for separation, black flags for negative fill, crisp specular control, and defined shadows with retained detail. Realistic moulded vinyl, premium cardboard, clear PET window, accurate seams and inserts, lifelike materials, precise reflections, magazine-quality retouching, soft bloom, subtle grain, clean white product backdrop, anatomically correct hands, no extra fingers, no warped anatomy, no plastic skin, no blown highlights, controlled whites, retained shadow detail, no logo, no watermark, no readable signage.

Character Rebuilt from Sliced Fruits and Vegetables

game-asset

A detailed prompt for Nano Banana Pro designed to convert an uploaded character reference into a realistic 3D render made entirely of layered, sliced fruits and vegetables, maintaining the original character's proportions, colors, and identity while ensuring all materials, including footwear, are constructed from produce slices.

展开提示词全文
create a version of this character made entirely of sliced vegetables and fruits vegetables and fruits: should match the uploaded character’s colors as closely as possible, without changing their natural look slices: slightly misaligned, with visible separation between layers structure: a small bite taken out of the top of the head, forming a clean cut surface ambience: kitchen setting, on a cutting board with scattered vegetables and a knife nearby camera: slightly top-down angle effect: tilt-shift  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 Character Compatibility Rules: If the uploaded character is 2D, convert it into a believable realistic 3D render while preserving the original design, proportions, colors, and visual identity Do not modify facial structure  Do not add a mouth or nose if the uploaded character does not have them  👟 FOOTWEAR & ACCESSORY MATERIAL OVERRIDE (CRITICAL FIX) All parts of the character, including footwear and clothing, must follow the same fruit and vegetable slice construction. • shoes must be made entirely of stacked fruit and vegetable slices  • preserve the exact shoe shape, proportions, and design  • recreate panels, sole, and details using layered slices Material examples for shoes: • sole → dense stacked potato or root vegetable slices  • upper panels → fruit/vegetable slices arranged to match original color blocking  • edges and seams → visible slice layering and separation 📷⚠️ 0px; letter-spacing: normal; orphans: 2; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px;"> IMPORTANT:  No original materials should remain (no fabric, rubber, or leather).  Everything must be converted into fruit/vegetable slice construction while keeping the original design recognizable. ar 1:1

Pixar-Style 3D Character Sticker Set (9 Variations)

game-asset

A detailed prompt for Nano Banana 2 to generate a 3x3 grid of 9 distinct Pixar-style 3D character stickers based on an uploaded facial reference image. The prompt emphasizes maintaining high likeness accuracy, specific framing (head and shoulders), varied modern outfits, and nine different expressions, all rendered in ultra-high resolution with a premium, glossy finish.

展开提示词全文
Create a premium Pixar-style 3D character sticker set that is clearly recognizable as the same person. Use the uploaded image as the ONLY facial reference.
Output Requirements:
3×3 grid layout (9 stickers total)
Final aspect ratio: 4:5
Ultra high resolution (8K)
Clean cream background
Each sticker is separated with clear spacing
Add a subtle drop shadow under each sticker for depth
Character Identity (VERY IMPORTANT):
MUST preserve exact facial structure, hairstyle, grooming details, and unique features from the uploaded image
Maintain high likeness accuracy
Do NOT over beautify or alter identity beyond recognition
Style:
Pixar-style 3D animation
Glossy, high-end rendering
Soft global illumination
Smooth shading
High-detail facial textures
Framing:
All stickers must be head and shoulders (portrait close-up) only
No full-body or mid-body shots allowed
Focus tightly on face and upper shoulders to emphasize expressions
Outfit:
Modern casual and slightly stylized clothing
Each sticker must have a different outfit (no repetition at all)
Clean, minimal, and visually appealing for sticker use
Expressions & Poses (9 variations):
1. Happy smile
2. Laughing
3. Cool (with sunglasses)
4. Angry/serious
5. Surprised
6. Thinking
7. Wink
8. Love (heart eyes and heart gesture)
9. Thumbs up
Sticker Design:
Bold, clean outline
Slightly exaggerated cartoon proportions
Strong emphasis on facial expressions
Optimized for digital sticker platforms
Each sticker must have a smooth rounded bottom cutout (semi-curved base)
No text
Lighting:
Soft studio lighting
Subtle rim light for depth and separation
Goal:
Create a highly engaging, cute, and expressive sticker pack that feels premium and collectable.

KNIGHT MECHA GIRL Figure Generation Prompt

game-asset

A prompt used to generate an image of a 1/7 scale figure of 'KNIGHT MECHA GIRL Re : No.3 La Flamme Blanche,' with a specific request for the back and the underside of the shield to be detailed.

展开提示词全文
KNIGHT MECHA GIRL
Re : No.3 La Flamme Blanche
⠀
1/7 scale figure
⠀
I would be happy if the back and the underside of the shield look realistic.

Reference Character Posing Grid Study Prompt

game-asset

A detailed prompt for nano banana to generate a 2x3 grid perspective study of a reference character, showing six specific panels from extreme wide to extreme close-up, focusing on maintaining character consistency and tailoring poses to the camera distance.

展开提示词全文
16:9 landscape, 2x3 grid perspective study of the reference character on a solid off-white background. 6 panels in strict order: 1. Extreme Wide (silhouette focus), 2. Full Body (dynamic pose), 3. Waist-Up (torso twist), 4. Bust Shot (strong posture), 5. Face Close-Up (expression), 6. Extreme Close-Up (facial details/eyes).
- Consistency: Maintain identical facial structure, hair, and proportions across all panels.
- Posing: Tailor gestures to camera distance—broad movements for wide shots, subtle nuances for close-ups.
- Aesthetic: Consistent lighting and art style; no borders, text, or background environments.

3D Model Asset Style: Classic Japanese Bosozoku (Female)

game-asset

A prompt used with Google Gemini 3.1 image generator (Nano Banana Pro) to create an image of a classic Japanese Bosozoku (female) in the style of a 3D model asset.

展开提示词全文
the classic Japanese Bosozoku (F) in the style of a 3D model asset

Nano Banana Pro Image Generation of Classic Japanese Bosozoku as a 3D Model Asset

game-asset

A prompt used with Nano Banana Pro (Google Gemini 3.1 image generator) to create an image of a classic Japanese Bosozoku (male) rendered in the style of a 3D model asset. This is presented as an example of generating AI assets/materials.

展开提示词全文
the classic Japanese Bosozoku ({argument name="gender" default="M"}) in the style of a 3D model asset

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。

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。

AI Image Prompts · 绘画提示词库

官方标注免费

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

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。