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 2105 条
Chibi Character Tehepero Pose Prompt for Nano Banana Pro
profile-avatarA detailed prompt for Nano Banana Pro to generate a super-deformed (chibi) character performing the 'Tehepero' pose (sticking out the tongue and winking). The prompt specifies the character's action, gaze, camera angle, perspective, lens, and composition, emphasizing a big head and small body proportion, and requires an attached image for the character reference.
展开提示词全文
The chibi character in the attached image is performing the Tehepero pose. She is making a fist with one hand and placing it on her temple. Her tongue is sticking out upwards from the corner of her mouth. She is winking with one eye. Gaze: soft direct gaze, looking at the viewer, gentle eye contact Camera angle: high angle (not overhead), looking down at the subject, face visible Perspective: exaggerated perspective, strong foreshortening Lens: wide-angle (24mm), close to camera Composition: full body, centered, simple white background Proportions: big head, small body (super-deformed)
Chibi Character 'Tehepero' Pose Generation Prompt
profile-avatarA detailed structured prompt for Nano Banana Pro to generate a chibi character performing the 'Tehepero' pose (winking and sticking out the tongue slightly). It specifies facial expressions, body pose, camera angle (high angle, 24mm equivalent), and composition (full body, centered, simple white background). This prompt requires an existing character image as input.
展开提示词全文
subject:
description: "The chibi character in the attached image is performing the 'Tehepero' pose"
character_style:
- "chibi character"
expression_and_pose:
facial_expression:
- "winking the right eye"
- "tongue slightly sticking out upwards from the corner of the mouth ('Tehepero')"
pose:
- "lightly tapping the high side of the head with the right fist"
- "left arm is stretched straight downwards"
body_orientation:
torso:
- "The front of the body is not completely facing the camera, but slightly angled (3/4 view)"
gaze:
direction: "looking at the camera"
quality:
- "soft frontal gaze"
- "gazing at the viewer"
- "gentle eye contact"
camera_and_composition:
angle:
- "high angle"
- "looking down from above, but the face is clearly visible"
perspective:
- "exaggerated perspective"
- "strong foreshortening (foreground appears large, background appears small)"
lens:
focal_length_equivalent: "24mm equivalent"
distance: "camera is close to the subject"
framing:
- "full body included"
- "centered"
background:
- "simple white background"
Raw High-Definition Printed Portrait Aesthetic Prompt
profile-avatarA highly detailed prompt designed to emulate a raw, high-definition photograph of a physical printed portrait, capturing the texture of the paper and ink. It focuses heavily on realism, specific lighting (harsh studio flash), and includes a creative text overlay, ideal for generating yearbook or vintage magazine archive aesthetics.
展开提示词全文
(3:4). A raw, high-definition photograph of a physical printed portrait taken with an iPhone 17 Pro. NO ARTIFICIAL BOKEH. Everything is SHARP and in focus, capturing the physical textural reality of the paper grain and the ink-on-page in high resolution.
Subject & Identity Locking:
Identity: Strictly preserve the exact face and unique features of the woman in Image 1 (Selfie).
Bio-Fidelity: Render high-fidelity "TrueLens" skin physics as captured through high-end printed ink: visible micro-pores, natural "satin-finish" glow muted by paper texture, and fine peach fuzz along the jawline.
Expression: A perfectly aligned, symmetrical head-and-shoulders yearbook-style pose. Neutral but confident expression with a magnetic, "main character" gaze looking directly into the lens.
Silhouette Calibration: Despite the formal school tailoring, amplify the dominant structural bust volume and the aggressive hourglass structural tapering at the waist. Ensure the fit of the uniform highlights a high-impact, fit physique.
Outfit & Material Realism:
Clothing: A dark navy blue structured blazer with white piping, a crisp white collared shirt, and a dark navy striped silk tie.
Tension: The blazer and shirt show extreme material tension and a vacuum-tight fit, creating realistic physical interaction and "tension lines" that highlight her fit, symmetrical upper silhouette and anatomical depth.
The Printed Aesthetic (Hardware Emulation):
Surface Texture: The image features an authentic semi-gloss paper sheen with thin, realistic diagonal lamination glares reflecting off the ink surface across the cheeks and chest.
Ink Physics: Visible ink-absorption softening at the edges (not digital blur), fine paper grain texture, and a gentle warm undertone typical of vintage magazine archives. It looks like a scanned physical object.
Lighting & Composition:
Lighting: Direct frontal harsh studio flash (ID style). Even exposure with bright, sharp eye catchlights and minimal shadows, creating a flat but clean professional portrait look.
Background (SHARP): A flat, textured cool-blue studio backdrop, showing realistic ink-on-paper depth.
Creative "Pinterest" Text Overlay:
Center-aligned below the portrait in an ink-printed bold serif font: "{argument name="main text" default="MOST LIKELY TO..."}"
Below in smaller, elegant italic subtext: "{argument name="subtext" default="“Be the reason you check your phone every 5 minutes.”"}"
The text features slightly imperfect ink bleeding and natural letter-spacing, mimicking an authentic high-end physical printing process.
Quality: Raw scan/photo-of-print aesthetic, high dynamic range, visible paper fiber and digital ISO noise in the shadows, 8k resolution, looks physically touchable.
Pixar 3D Character Portrait with Fennec Fox
profile-avatarA prompt for generating a cute, cozy 3D Pixar-style character portrait of a little girl with a fennec fox resting on her shoulder, emphasizing soft smooth skin, detailed knitted textures, and a warm color palette.
展开提示词全文
1. A cute 3D Pixar-style character portrait of a little girl with big expressive eyes and round glasses, soft smooth skin, and slightly rosy cheeks. She has dark hair styled into two messy space buns tied with red scrunchies. She is wearing a cozy red knitted sweater with white winter patterns and blue denim overalls with small fabric details and light wear. Her arms are crossed casually. A small adorable fennec fox with huge ears and sleepy eyes is resting on her shoulder, leaning against her neck affectionately.
2. A cute 3D Pixar-style character portrait of a young girl with big round glasses and twin messy hair buns tied with red bands, wearing a red knit sweater under light blue denim overalls with small patches and worn details. She is standing with her arms crossed, smiling gently. A tiny sleepy fennec fox is resting on her shoulder. The background is a clean solid {argument name="background color" default="coral-red"} color. Soft studio lighting, shallow depth of field, ultra-detailed, high quality render, smooth skin, warm color palette, cozy and wholesome mood, centered composition, 4k, cinematic lighting.
Vintage Film Korean Idol Mirror Selfie Collage
profile-avatarA complex, multi-panel prompt designed to replicate a vintage film aesthetic (Kodak Portra/Cinestill) using slow shutter sync flash (shutter drag) for a dreamy, blurred, intimate 2x2 mirror selfie collage of a Korean female idol in her bedroom.
展开提示词全文
{
"global_style_injection": "Heavy film grain, soft focus, dreamy Pro-Mist diffusion filter, motion blur from camera panning, vintage color grading, flash photography",
"subject": {
"description": "A 2x2 mirror-selfie collage of a charming Korean female idol in her bedroom. The image features a distinct 'shutter drag' effect where the subject is illuminated by flash but slightly blurred due to movement, creating a dreamy, hazy aesthetic.",
"vibes": "Sexy yet innocent, intimate, slightly intoxicated mood (drunk on atmosphere), atmospheric, lo-fi aesthetic.",
"looks": {
"face": "Soft features, glowing skin due to diffusion filter, 'water-color' like makeup texture",
"eyes": "Dreamy gaze, slightly out of focus in one panel, looking directly into the lens with a soft smile in another",
"hair": "Long black hair, moving with the motion blur, creating silky trails"
}
},
"clothing": {
"style": "Pure and sexy loungewear",
"item": "A white silk slip dress or a translucent oversized white t-shirt falling off one shoulder",
"texture": "Silk or soft cotton that glows under the flash and diffusion filter",
"details": "Hint of collarbone, casual disarray"
},
"pose": {
"action": "Playful and dynamic movement. Not static.",
"panels": [
"Panel 1: Laughing while turning head, hair in motion blur",
"Panel 2: Close-up wink with heavy Pro-Mist bloom on the face",
"Panel 3: Sitting on floor, camera tilted, slight double-exposure effect from movement",
"Panel 4: Covering mouth with hand, shy but flirting"
]
},
"background": {
"setting": "Personal bedroom, slightly messy but cozy",
"lighting_elements": "Warm bedside lamp, fairy lights, or neon sign reflection that creates light trails (bokeh streaks) due to the camera panning motion",
"ambiance": "Dark room illuminated by the harsh but diffused flash"
},
"photography": {
"technique": "Slow shutter sync flash (Drag Shutter)",
"filter": "{argument name="diffusion filter" default="Black Pro-Mist 1/4 filter"} (creates halation around highlights)",
"focus": "Soft focus, slightly missed focus for artistic effect",
"color_grading": "Vintage film stock ({argument name="film stock" default="Kodak Portra 800 or Cinestill 800T"}), greenish shadows, warm glowing highlights",
"texture": "Heavy film grain, noise, imperfection, analog feel"
},
"constraints": {
"layout": "2x2 grid collage",
"composition": "The camera panning effect should make the background lights streak horizontally or rotationally, while the flash freezes the face relatively clearly."
}
}
Low-Resolution Motion-Blurred Selfie with Cat
profile-avatarA highly specific negative prompt designed to generate a raw, imperfect, low-resolution smartphone selfie aesthetic, featuring a woman and a cat with natural, uneven motion blur caused by poor camera quality and movement, explicitly forbidding high-quality features like sharp focus or AI enhancement.
展开提示词全文
Use the uploaded reference photo. Low-resolution smartphone close-up selfie taken with a broken front camera screen. The image shows strong natural motion blur concentrated on the woman’s face, hair, and the cat, caused by weak autofocus, slow shutter speed, and slight hand movement while lying down. Blur appears uneven and organic, with soft smearing rather than smooth artificial blur. Facial features are partially obscured but still recognizable, with eyes, nose, and mouth visible as softened shapes. A young woman lies intimately on a cream-colored leather sofa. A light brown cat with bright green eyes is lying on the woman’s chest, pressing its body close and nuzzling its face against the woman’s cheek and jaw, partially covering her face. The cat’s head and the woman’s face overlap closely, creating intimate contact. Both the woman’s face and the cat’s face are noticeably blurred together, with softened edges and slight motion smear from small movements and breathing. The woman’s long brown hair is loose and slightly messy, falling over her forehead and partially blending into the cat’s fur due to blur. Hair strands appear heavily softened but still natural, not melted or warped. The cat’s fur texture is mostly lost, appearing as soft blurred patches consistent with low resolution and motion. Overall sharpness is low, with inconsistent focus across the frame — some tiny areas accidentally a bit clearer while most of the face and the cat remain out of focus. Flat, noisy skin texture with visible ISO noise, chroma noise, mild color banding, and muddy shadows typical of a cheap phone camera. The woman looks toward the lens with a soft, relaxed, intimate expression, though details are softened by blur. She wears simple, light-colored casual clothing, with fabric texture mostly lost. Lighting comes only from indoor room lights at night, slightly too bright, with mild overexposed highlights that do not completely blow out facial structure. Perspective distortion from the front camera wide-angle lens creates subtle stretching near the edges. The composition is very tight, awkward, and unintentional, filling the frame with the woman’s face and the cat’s body and head. The image feels like a raw, accidental selfie straight from the phone gallery — imperfect, intimate, noisy, naturally blurred, completely unedited, with no beautification and no AI enhancement. NEGATIVE WARNINGS (important): high resolution, sharp focus, DSLR, cinematic lighting, studio lighting, ultra-detail, clean skin texture, beauty filter, HDR, professional photo, sharp edges, smooth look, AI enhancement, overprocessing, perfect blur, depth-of-field bokeh. ar 9:16
Raw Wide-Angle Glitter Dress Selfie
profile-avatarA highly technical prompt for generating a raw, high-contrast, wide-angle smartphone selfie aesthetic, emphasizing hyper-realistic skin physics, aggressive body contouring, and harsh flash lighting, with a critical constraint of zero artificial bokeh and sharp focus throughout the entire frame.
展开提示词全文
(Vertical 9:16 smartphone aspect ratio). A raw, high-contrast handheld 0.5x wide-angle selfie captured with an iPhone 17 Pro. NO ARTIFICIAL BOKEH. Everything is SHARP and in focus from her glossy lips to the fabric of the sofa.
Subject & Identity Locking:
Identity: Strictly preserve the exact face and unique features of the woman in Image 1 (Selfie).
Bio-Fidelity: Render high-fidelity skin physics: visible micro-pores, natural "satin-finish" night hydration, and ultra-glossy, plump lips. Fine peach fuzz is visible along the jawline.
Expression: A confident, heavy-lidded seductive gaze with a soft smirk directed at the lens.
Silhouette Calibration: Prioritize a dominant structural bust volume and substantial, thick-toned leg presence. Focus on the aggressive hourglass structural tapering at the waist.
Outfit & Physicality:
Clothing: A skin-tight, {argument name="dress color" default="Champagne-Gold"} Glittery Strapless Mini Dress.
Tension: The shimmery, reflective fabric shows extreme elastic tension and vacuum-tight fit, creating realistic micro-folds at the waist and highlighting her fit, symmetrical proportions.
Accessories: A delicate gold chain necklace with a small blue and white "Evil Eye" (Nazar) pendant resting on her collarbone.
Physics: Realistic skin-to-surface compression (thigh squish) is visible where her legs meet the sofa.
Pose & Composition:
Pose: She is reclining back into a sofa. High-angle downward POV (0.5x) looking at her torso and legs. Her body is slightly twisted to maximize the S-curve silhouette.
Action: Her RIGHT ARM is visible in the frame, holding the phone high for the selfie.
Environment & Lighting:
Lighting: Direct, harsh indoor lighting (mimicking a smartphone flash). The light creates brilliant starburst reflections on the glittery dress and a high-key glow on her skin.
Background (SHARP): A detailed domestic setting. A grey textured fabric cushion and the beige leather-grain of the sofa back are all perfectly IN FOCUS. No artificial blur.
Quality: Raw unedited "fit-check" social-media dump aesthetic, high dynamic range, visible digital grain in the deep shadows, 8k resolution.
Ultra-Realistic Cinematic Night Street Selfie Prompt
profile-avatarA highly detailed, complex prompt designed for Nano Banana PRO to create an ultra-realistic cinematic night street selfie. It emphasizes face locking for identity preservation, dynamic lighting (police sirens, streetlights), specific fashion details (leopard print coat, martini glass), and a rebellious, glamorous chaos mood, suggesting a high-speed escape scenario.
展开提示词全文
Using my uploaded photo as the sole face reference, create an ultra-realistic cinematic night street selfie. Preserve the original facial identity 100% with no changes to facial structure, proportions, or unique features (face locked, identity preserved).
A confident, stylish woman leaning out of a car window in a rainy city during a high-speed escape.
Pose: arm extended toward the camera holding the phone, head slightly tilted back, tongue playfully liftedtoward the upper lip in a teasing, provocative gesture; fearless, playful attitude.
Face & Skin: smooth, luminous skin with glossy streetlight highlights, softly sculpted cheekbones, glossy lips slightly parted, expressive eyes partially hidden behind narrow black rectangular sunglasses; realistic skin texture, pores, natural reflections.
Hair: long, flowing hair with natural volume, wind-blown by speed and night air.
Outfit & Details: bold {argument name="coat type" default="leopard-print faux-fur coat"} draped luxuriously overthe shoulders, glossy red manicure, minimal silver rings, elegant martini glass in hand with a clear cocktail and green olives catching sharp reflections.
Environment: wet asphalt night city street, intense motion; police car chasing behind with flashing red and blue sirens; blurred headlights, aggressive light streaks, cinematic urban depth; golden street lamps creating large dramatic bokeh circles mixed with police light flares; reflections stretching across the road.
Lighting: dynamic mixof warm sodium streetlights, cold blue police lights, and red siren flashes; high contrast; glossy reflections on skin, glass, car body, and rain-soaked street.
Mood: rebellious, playful, glamorous chaos; adrenaline-fueled nightlife energy; fashion-editorial crime aesthetic.
Camera & Style: wide-angle smartphone selfie perspective, shallow depth of field, strong background motion blur, light streaks, ultra-high detail, realistic film grain, cinematic color grading, modern luxury runaway vibe.
Photorealistic Portrait of a Korean Young Woman in a Gamer Chair
profile-avatarA detailed, photorealistic prompt for generating a high-quality portrait of a Korean young woman. The prompt specifies visual details like overwhite skin, dewy look, blue eyes, specific clothing (pink fuzzy oversized sweater, black-framed glasses), pose (sitting with crossed legs, making an 'L' sign), setting (gamer chair, laptop, soft grey background with hexagonal patterned dolls), and technical specifications (8K resolution, cinematic lighting, 9:16 aspect ratio).
展开提示词全文
A high-quality, photorealistic portrait of a {argument name="subject" default="Korean young woman"} with overwhite skin, dewy look, blue eyes with long lashes, and nude gradient pink glossy lips. She's sitting comfortably with crossed legs on a gamer chair, with a laptop in front of her. She's wearing oversized thick black-framed glasses and a cozy {argument name="sweater color" default="pink"} fuzzy oversized sweater. Her hair is styled in a messy high bun. She's looking at the camera with sparkling eyes and making an 'L' sign with both hands. Warm indoor lighting, soft grey background with hexagonal patterned dolls, aesthetic 'clean girl' vibe, 8K resolution, cinematic lighting, aspect ratio 9:16. This is not a cartoon or painting. 100% copy of my face (no hijab).
Cinematic Portrait Prompt for Outdoor Expedition Vibe
profile-avatarA detailed, structured prompt designed for the Gemini App (Nano Banana Pro) to generate a hyper-realistic, cinematic portrait of a male subject in a rugged outdoor setting. The prompt specifies every visual detail, including the subject's features, clothing, mountain scene, dramatic lighting, and precise camera settings (85mm lens, shallow depth of field) to achieve a high-end professional photography aesthetic.
展开提示词全文
{
"goal": "Produce a cinematic, lifelike portrait with a bold outdoor expedition vibe",
"subject": {
"identity": "same person as the reference — identical face and proportions",
"gender": "{argument name="gender" default="male"}",
"age": "{argument name="age" default="late 20s"}",
"look": {
"hair": "messy, wind-blown texture",
"beard": "trimmed and short",
"features": "defined, angular structure",
"mood": "focused and intense"
},
"skin": "high-fidelity detail with natural color variation and realistic pores"
},
"clothing": {
"jacket": "{argument name="jacket style" default="distressed dark leather coat"}",
"notes": "visible wear, tactile rugged surface, reinforced stitching"
},
"scene": {
"setting": "remote mountain terrain",
"background": {
"peaks": "soft layers of fog across distant mountains",
"forest": "evergreens dissolving into haze",
"air": "thick, moody mist"
}
},
"lighting": {
"style": "cinematic and atmospheric",
"source": "diffused daylight from cloudy skies",
"shadows": "natural, dramatic shaping",
"palette": "cool, slightly desaturated colors"
},
"camera": {
"lens": "85mm portrait lens",
"aperture": "f/1.8",
"dof": "very shallow depth of field",
"focus": "precise on eyes and face, smooth background blur"
},
"aesthetic": {
"render_quality": "high-end professional photography",
"realism": "hyper-realistic",
"finish": "fine micro-textures, sharp clarity, cinematic tone curve"
}
}
Hyper-realistic mirror selfie prompt with detailed material physics and identity locking
profile-avatarA highly detailed image generation prompt designed to create an extremely realistic, raw, high-contrast mirror selfie. It specifies strict identity locking, bio-fidelity for skin textures (TrueLens physics), precise outfit material realism (vacuum-tight jersey knit), and environmental details like realistic mirror imperfections and a sharp background, aiming for an unedited social media dump aesthetic.
展开提示词全文
(Vertical 9:16 smartphone aspect ratio). A raw, high-contrast mirror selfie captured with an iPhone 17 Pro. NO ARTIFICIAL BOKEH. Everything is SHARP and in focus from the ornate mirror frame to the wooden textures in the background. Subject & Identity Locking: Identity: Strictly preserve the exact face and unique features of the woman in Image 1 (Selfie). Bio-Fidelity: Render high-fidelity "TrueLens" skin physics: visible micro-pores, a natural "satin-finish" glow, and fine peach fuzz catching the light ONLY along the jawline. Expression: Looking into the mirror with a confident, calm gaze and neutral, full lips. Silhouette Calibration: Prioritize a dominant structural bust volume and an aggressive hourglass structural tapering at the waist. Maintain a symmetrical and fit anatomical depth. Outfit & Material Realism: Clothing: A skin-tight, jet-black jersey-knit long-sleeved dress. The neckline is styled asymmetrical, sliding off her RIGHT shoulder to reveal her collarbone. Tension: The fabric shows extreme elastic tension and a vacuum-tight fit, creating realistic micro-folds and "tension lines" that highlight her fit silhouette and the physical interaction between the textile and her form. Accessories: A delicate silver chain necklace with interlocking double-circle pendants resting on her décolletage. Pose & Composition: Pose: A classic indoor mirror selfie. She is standing with a slight hip-pop to emphasize the S-curve of her physique. Action: Her LEFT HAND holds an iPhone 17 Pro (with a prominent circular grip/PopSocket on the back) at chest level, capturing the reflection. Environment & Social Realism: Mirror: Captured through an ornate octagonal gold-gilt framed mirror. Crucially, the glass features realistic imperfections: faint water spots, dust particles, and minor smudges that catch the light for a "raw evidence" look. Background (SHARP): A domestic bathroom setting. Behind her, a dark wood-grain door and a neutral-toned wall are perfectly IN FOCUS. The chrome handle of a sink faucet is visible at the bottom of the frame. Quality: Raw unedited social-media dump aesthetic, high dynamic range, visible digital grain in the deep shadows, 8k resolution.
Cinematic Winter Forest Portrait Prompt
profile-avatarA detailed, structured prompt designed for generating a highly photorealistic and cinematic portrait of a young woman in a snowy winter forest setting. The prompt specifies persona details, scene description, photography style, and art direction, including a negative prompt for quality control.
展开提示词全文
{
"Objective": "Generate a cinematic, photorealistic portrait inspired by a winter forest atmosphere",
"PersonaDetails": {
"Subject": "{argument name="subject" default="Young woman"}",
"Appearance": {
"Hair": "Long, wavy dark brown hair, slightly windblown",
"Eyes": "Light blue",
"Skin": "Fair with natural freckles",
"Expression": "Soft, introspective, slightly wistful"
},
"Clothing": {
"Outerwear": "Dark brown wool coat",
"Accessories": "Knitted scarf in neutral tones"
}
},
"SceneDescription": {
"Environment": "{argument name="environment" default="Snowy forest with soft-focus evergreen trees"}",
"Weather": "Gentle snowfall",
"Season": "Winter",
"Mood": "Quiet, cinematic, melancholic, ethereal"
},
"PhotographyStyle": {
"Type": "Cinematic portrait photography",
"Lighting": "Soft natural light, overcast winter lighting",
"DepthOfField": "Shallow depth of field with creamy bokeh",
"CameraAngle": "Three-quarter profile, subject looking back over shoulder",
"Resolution": "Ultra high resolution, sharp facial details"
},
"ArtDirection": {
"ColorPalette": "Muted earth tones, cool greens and soft whites",
"Texture": "Natural skin texture, visible snow particles",
"RealismLevel": "Highly photorealistic"
},
"NegativePrompt": [
"cartoon",
"anime",
"overly smooth skin",
"harsh lighting",
"blurry face",
"distorted anatomy",
"oversaturated colors"
],
"ResponseFormat": {
"Type": "Single image",
"Orientation": "Portrait",
"AspectRatio": "2:3"
}
}
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)
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)
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。
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,在海报排版上搭配清晰视觉层级、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。
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。