Ai Image Prompts
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
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。
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
Disney-Pixar Style Hybrid Sloth-Dragon Creature
game-assetA structured prompt for generating a high-fidelity 3D animation image in the Disney-Pixar style. It describes a cute hybrid creature (sloth fur and dragon scales) sitting on a mossy log and holding a glowing, magical pink orchid in a sun-drenched tropical jungle.
展开提示词全文
{
"image_generation_prompt": {
"subject": "A cute, wide-eyed baby hybrid creature with a mix of soft, fluffy brown sloth fur and iridescent green dragon scales, featuring small orange wings.",
"action": "The creature is sitting on a mossy log, holding a glowing, magical pink orchid flower that emits a soft, ethereal light.",
"environment": "A lush, sun-drenched tropical jungle canopy with dense foliage, hanging vines, and soft bokeh.",
"art_style": "High-fidelity 3D animation, Disney-Pixar style, featuring large expressive eyes and detailed textures.",
"lighting": "Cinematic lighting with natural sunlight filtering through leaves (God rays) combined with a magical bioluminescent glow from the orchid.",
"composition": "Close-up shot focusing on the creature's endearing expression and the contrast between fur and scale textures."
},
"metadata": {
"aspect_ratio": "16:9",
"quality": "Ultra-HD",
"style_reference": "Modern 3D Feature Film"
}
}
Hyperrealistic Tattooed Elf Couple
game-assetA prompt for a hyperrealistic portrait of a beautiful, tattooed elf and warrior couple set in a winter scene with snow and a castle background. The female elf is specified to have a sword, blue eyes, long white hair, and knightly armor.
展开提示词全文
hyperrealistic portrait of a beautiful, tattooed elf and warrior couple. Winter, snow, a woman holding a sword. A castle in the background. The elf woman has a tattoo, blue eyes, long white hair, and knightly armor.
Mech Armor Power Reveal Prompt (Likeness Preservation)
game-assetA highly technical prompt for generating a cinematic power reveal of a character fully re-engineered in advanced mech armor, with critical instructions to preserve the exact original facial geometry, eye shape, and color palette, using a worm's-eye ultra-low angle.
展开提示词全文
An ultra-dominant cinematic power reveal of the character fully re-engineered in advanced mech armor, preserving the exact eye shape, spacing, facial structure, patterns, and original color palette with absolute fidelity. The helmet is open, framing the face without altering proportions. Identity remains untouched. No humanization. No anatomical reinterpretation. Likeness Preservation (CRITICAL): Keep the exact eyes, exact facial geometry, exact stylization. Armor integrates around the character’s existing design language. Patterns are embedded into armor plating, not replaced. Colors remain identical, only expressed through reinforced materials and illuminated seams. Perspective: Worm’s-eye ultra-low angle shot, camera almost touching the ground beneath the character. 16–18mm wide lens. Slight 7–10 degree horizon tilt to destabilize symmetry. One armored limb closer to the lens, dominating foreground mass. Composition: The character occupies 70% vertical frame height. Helmet open. Face near upper third intersection point. One massive mech gauntlet partially cropped out of frame edge to create unsafe framing tension. Pose: Not static. Transitional power ignition. One step forward mid-weight shift. Shoulder rotation engaged. Armor plates subtly separating as internal energy surges beneath. Depth of Field Discipline: Primary focal plane locked on the eyes. Foreground mech limb slightly soft. Lower legs progressively fall into depth blur. Background collapses into volumetric haze. Lighting Structure: Hard directional back cross-light slicing through armor edges. Cold blue rim accent along shadow edge. Subtle warm internal glow leaking through mech seams. Deep shadow pockets between plating layers. Specular highlights sharp and controlled, not glossy plastic. Surface Treatment: Micro-scratches on armor. Slight burnished metal variation. Energy veins glowing beneath translucent panels. Face remains clean and consistent with original material identity. Environment: Minimal but monumental. Volumetric atmosphere. Light shafts cutting diagonally behind character. Ground plane textured but subdued, darker than subject. Emotion: Contained power. Controlled destruction. The mech is not worn. It has bonded. Negative Prompt: No redesign of eyes. No change in facial proportions. No color shift. No bulky random sci-fi accessories. No symmetrical hero pose. No flat studio lighting. Aspect Ratio: ar 4:5
Detailed Character Prompt: Vampire Girl with Werewolves
game-assetA highly structured JSON prompt for generating a cinematic fantasy image of a vampire girl sitting on a log in a dark forest, flanked by three highly detailed werewolf creatures, focusing on gothic Victorian style and moody lighting.
展开提示词全文
{
"scene": "A young woman dressed as a vampire sitting on a mossy log in a dark forest, flanked by three realistic werewolf creatures.",
"subject": {
"character": "Vampire girl",
"face": {
"structure": "Symmetrical, youthful, soft features",
"skin": "Fair, pale complexion",
"eyes": {
"shape": "Almond-shaped",
"color": "Dark brown",
"expression": "Friendly, slightly mischievous smile"
},
"mouth": {
"lips": "Natural pink tone with visible vampire fangs and small droplets of blood at the corners"
}
},
"hair": {
"color": "Dark brown / Black",
"length": "Long, flowing behind shoulders"
}
},
"creatures": {
"type": "Werewolves",
"count": 3,
"appearance": "Highly detailed animatronic or costume style, thick brown/black fur, snarling expressions, visible yellowed fangs, pointed ears"
},
"pose": {
"overall": "Relaxed sitting position",
"position": {
"base": "Seated on a fallen moss-covered tree trunk",
"orientation": "Centrally framed, facing the camera"
}
},
"outfit": {
"dress": {
"type": "Gothic Victorian style gown",
"color": "Deep black",
"material": "Velvet body with lace detailing"
}
},
"environment": {
"location": "Dense, dark forest",
"ground": "Covered in dry brown leaves and a large mossy log"
},
"lighting": {
"type": "Natural, diffused woodland light",
"quality": {
"intensity": "Soft, moody"
}
},
"photography": {
"angle": "Eye-level",
"framing": "Full body shot, centered subject",
"style": "Cinematic fantasy"
},
"color_palette": {
"subject": {
"hair": "Black",
"skin": "Pale porcelain",
"dress": "Jet black"
},
"environment": {
"wood": "Deep brown and moss green",
"leaves": "Earth tones, copper, and tan",
"background":
Anime Stealth Scene with Keystoning Angle
game-assetA prompt for generating an anime-style image using a specific 2-point perspective and keystoning camera angle, depicting a tense stealth scene split compositionally between a hiding woman and a patrolling male guard.
展开提示词全文
Anime character in a background, 2-point perspective with a keystoning angle, composition split in half: on the right, a 20-year-old woman is pressed against a wall, waiting and hiding; on the left, seen through a door frame, a male guard holding a submachine gun is walking stealthily. Japanese style animation, full color, modern Madhouse style (2020s), detailed cel shading, cinematic post-processing, masterpiece.
Dynamic Video Prompt: Banana Float Rocket
game-assetA prompt for generating a dynamic video of a glowing banana-shaped float soaring forward like a rocket through space, leaving a sparkling cosmic dust trail, with the character cheerfully waving and the camera tracking smoothly from behind.
展开提示词全文
A glowing banana-shaped float soaring forward like a rocket through space, leaving a trail of tiny sparkling particles that resemble cosmic dust. The character cheerfully waves with pride while flying. The camera tracks smoothly from behind, following the motion, creating a dynamic cinematic feel. Add playful, lighthearted sci-fi sound effects to enhance the fun, adventurous atmosphere.
Anthropomorphizing LLMs
game-assetA prompt used to generate images of anthropomorphized LLMs (ChatGPT-5.2 think, Gemini 3.1 pro, Claude, Grok) with specific expressions and poses. The first image was generated by Nano Banana Pro, the second by ChatGPT.
展开提示词全文
Extremely anthropomorphize ChatGPT-5.2 think, Gemini 3.1 pro, Claude, and Grok, characterizing them with creative expressions and poses, and arranging them in an image.
Character Design Sheet for a Villainess
game-assetA prompt used with Nano Banana Pro to generate a character design sheet for a villainess, including specific facial expressions and suggestive dialogue, without explicitly using sensitive keywords.
展开提示词全文
In the character design sheet for a villainess, include a specific facial expression (without writing the specific sensitive word in the prompt) in the expression set, and make her say obscene lines.
Stylized Wicker Caricature Figurine
game-assetA prompt for generating a highly detailed, photorealistic image of a stylized caricature figurine made entirely from woven wicker/rattan material, emphasizing hyper-realistic texture, studio lighting, and artistic craftsmanship.
展开提示词全文
A highly detailed stylized caricature figurine of a person made entirely from woven wicker / rattan material, big head and small body proportions, hyper-realistic texture with intricate basket weave patterns forming facial features, hair, and clothing. The figure stands on a small woven base placed on a wooden table. Background is a warm wooden wall with subtle brand logos or props. Studio lighting, soft shadows, shallow depth of field, ultra-detailed, 8K resolution, sharp focus, photorealistic, artistic craftsmanship, handcrafted sculpture look.
Ornate Box Diorama of a Royal Treasury Counting Room
game-assetA complex, multi-layered master prompt for generating an ultra-realistic cinematic miniature diorama photograph of a Royal Treasury Counting Room, detailing strict geometric constraints, material variables, scene logic, and a hard front-right isometric camera system.
展开提示词全文
ORNATE BOX DIORAMA — TRI-LAYER MASTER PROMPT (HARD FRONT-RIGHT ISOMETRIC) THEME_LOCATION = Royal Treasury Counting Room BOX_VARIABLE = Silver DOOR_MATERIAL = Gold WINDOW_MATERIAL = Gold WINDOW_PANE_MATERIAL = FLOOR_MATERIAL = STAIRCASE_TYPE = Spiral STAIRCASE_MATERIAL = Silver MEZZANINE_FLOOR_MATERIAL = Silver DIORAMA_LIGHTING = ======================== LAYER 1 — PHYSICAL GEOMETRY (HIGHEST PRIORITY) The ornate antique box is a solid rigid container transformed into architecture. If BOX_VARIABLE is specified, the entire container and its carved architectural structure must be made from that material. If BOX_VARIABLE is empty, infer logically from THEME_LOCATION. The full outer shell must remain intact on all sides. No removed walls. No dollhouse cutaway. FRONT FACADE LOCK: The FRONT SIDE must contain: • exactly ONE centered door • exactly TWO windows • one window left of the door • one window right of the door Openings must be carved THROUGH the existing wall. Apply DOOR_MATERIAL / WINDOW_MATERIAL / WINDOW_PANE_MATERIAL if specified, otherwise infer logically. Front wall must remain structurally thick and complete. ORNAMENTAL DEPTH CONTINUITY LOCK: All carvings must continue through wall thickness. Frames must appear carved from the same solid material. Interior walls share identical structural thickness. The lid must match base width and length and be hinge-connected. ======================== LAYER 2 — SCENE LOGIC Create a two-level miniature THEME_LOCATION fully integrated into the container. Use FLOOR_MATERIAL / STAIRCASE_TYPE / STAIRCASE_MATERIAL / MEZZANINE_FLOOR_MATERIAL if specified, otherwise infer logically. Upper level attaches directly to container walls. The inside of the lid functions as the vaulted ceiling. All props, materials, and architecture must match THEME_LOCATION. ======================== LAYER 3 — RENDER & CAMERA Ultra-realistic cinematic miniature diorama photography. Use DIORAMA_LIGHTING if specified. ======================== HARD FRONT-RIGHT ISOMETRIC CAMERA SYSTEM CORNER ANCHOR RULE: The FRONT-RIGHT TOP CORNER of the box must be the closest point to the viewer. WALL VISIBILITY RATIO RULE: The front wall must occupy slightly more visible area than the right wall. Approximate visual dominance: • front wall ≈ 60% visibility • right wall ≈ 40% visibility Both must be clearly visible simultaneously. RIM OCCLUSION RULE: Camera must remain above the front wall rim. The top edge of the front wall must overlap a small portion of the interior floor. Interior must be seen OVER the wall, never through a missing wall. FORBIDDEN CAMERA POSITIONS: Do NOT center the camera on the front wall. Do NOT use a straight frontal view. Do NOT use a top-down view. Do NOT prioritize interior visibility over box geometry. COMPOSITION TARGET: The image must resemble a professionally photographed miniature collectible box viewed from a stable elevated front
Nano Banana Pro Prompt: Final Fantasy Leviathan vs Player (Cyberpunk)
game-assetA prompt for Nano Banana Pro (Gemini 3 Pro Image) to generate an image of the Final Fantasy summon Leviathan battling a player, rendered in a Cyberpunk style. The user then tried to increase the number of players to 9.
展开提示词全文
Final Fantasy Summoned Beast Leviathan VS Player in Cyberpunk style! ↑ Cyberpunk (template) Next Make the players 9 people...
Final Fantasy Leviathan VS Player and Sacred Geometry
game-assetA sequence of three prompts: the first requests a fantastical depiction of the Final Fantasy summon Leviathan battling a player, followed by prompts for Sacred Geometry and Sacred Geometry combined with Fractal Structure.
展开提示词全文
Final Fantasy
{argument name="summon name" default="Leviathan"} summon beast VS player in a fantastical style
Next
{argument name="theme 1" default="Sacred Geometry"}
Next
Sacred Geometry and {argument name="theme 2" default="Fractal Structure"}
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。