← 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 2105 条

Boho-Chic Ultra Photorealistic Portrait Prompt

profile-avatar

A highly detailed, structured prompt for an ultra-photorealistic portrait of a young woman with bohemian styling. It focuses heavily on facial details (freckles, green eyes, natural skin texture), intricate accessories (layered gold jewelry, hair cuffs), and a specific aesthetic (Coachella vibe, raw photography) with technical camera specifications (Sony A7R V, 35mm f/1.8).

展开提示词全文
{
  "prompt_type": "Ultra Photorealistic Portrait",
  "subject": {
    "demographics": "Young female, light skin tone",
    "facial_features": {
      "eyes": "Piercing green eyes, sharp focus, natural eyelashes",
      "skin": "Natural texture with visible pores, prominent freckles across nose and cheeks, rosy flushed cheeks, subsurface scattering",
      "lips": "Full, glossy pink lips",
      "expression": "Soft, alluring gaze, looking directly at camera, relaxed mouth"
    },
    "hair": {
      "style": "Long, dark, wavy, bohemian style",
      "details": " interwoven with colorful embroidery threads (blue, red, green), adorned with small gold cuffs, beads, and hair rings, loose strands framing face"
    }
  },
  "apparel": {
    "top": "Olive green textured crinkle fabric crop top, deep V-neck wrap style",
    "accessories": {
      "arms": "Gold snake-style arm cuffs on both upper biceps",
      "neck": "Layered jewelry: beaded choker with gemstones, thin gold chain with star and moon charms, longer gold pendant necklace",
      "held_item": "Pink plush doll purse/bag with brown hair and a smiling face held in lap"
    }
  },
  "pose": {
    "type": "Selfie angle",
    "perspective": "Close-up POV, slightly high angle, arms visible extending to hold camera"
  },
  "environment": {
    "background": "Neutral beige textured mesh screen or sunshade fabric",
    "context": "Outdoor or semi-outdoor shaded area, festival or beach club setting"
  },
  "lighting": {
    "type": "Soft natural daylight",
    "quality": "Diffused, even lighting, no harsh shadows, soft catchlights in eyes"
  },
  "technical_specs": {
    "camera": "Sony A7R V",
    "lens": "35mm f/1.8 lens",
    "resolution": "8k, UHD",
    "render_style": "Octane Render, Unreal Engine 5",
    "details": "Hyper-detailed, sharp focus on eyes and freckles, depth of field blurring the mesh background slightly, film grain, chromatic aberration",
    "aesthetic": "Boho-chic, Coachella vibe, summer aesthetic, raw photography"
  }
}

Studio Portrait of Young Man with Focused Lighting

profile-avatar

A concise, structured prompt for Nano Banana Pro on the Gemini App, requesting a photorealistic studio portrait of a young man based on a reference image. The prompt emphasizes minimal, focused key lighting, a blue studio wall background, and a clean, cinematic aesthetic using an 85mm lens and shallow depth of field.

展开提示词全文
{
  "prompt": "Studio portrait of a young man, same appearance as the reference, posing confidently with relaxed body language. Minimal studio lighting, soft directional light fully focused on him while background remains subtle. He is wearing a {argument name="outfit" default="stylish casual outfit (casual shirt or tee, fitted jeans, clean sneakers)"}. {argument name="background color" default="Blue"} studio wall background, clean professional look, high-resolution, sharp details, cinematic yet natural, DSLR depth of field, realistic textures.",
  "style": "photorealistic",
  "lighting": "soft minimal focused key light",
  "camera": "85mm lens, shallow depth of field",
  "resolution": "4k",
  "mood": "confident, modern, composed"
}

Avatar Zuko Fan Photo Prompt for Nano Banana Pro

profile-avatar

A structured JSON prompt for Nano Banana Pro, instructing the generation of a hyper-realistic image of a subject posing with a giant 2D Zuko character from Avatar: The Last Airbender.

展开提示词全文
[
  {
    "image_generation": {
      "quality": "hyper-realistic",
      "face": { "preserve_original": true },
      "subject": {
        "clothing": "{argument name="clothing" default="dark red knitted sweater, dark brown sweatpants with yellow vertical lines on the sides, brown high-top sneakers"}",
        "pose": "standing with arm around giant 2D Zuco from avatar",
        "expression": "fun, mischievous",
        "character_element": {
          "name": "Zuco",
          "type": "3D & 2D photorealistic duo",
          "interaction": "Zuco posing confidently"
        }
      },
      "environment": "clean light green backdrop"
    }
  }]

Glamorous Mirror Selfie in Dior Oblique Top and Low-Rise Jeans

profile-avatar

A JSON prompt for generating a standing mirror selfie of a glamorous young woman, detailing her platinum blonde hair, full glam makeup, specific accessories (Louis Vuitton bag, iPhone Pro), and outfit (Dior Oblique scarf top and low-rise jeans) in a brightly lit, minimalist indoor setting.

展开提示词全文
{
  "subject": {
    "demographics": "Young female",
    "physique": "Slim, toned, tan skin",
    "pose": "Standing mirror selfie, holding phone with right hand, facing forward"
  },
  "appearance": {
    "hair": {
      "color": "Platinum blonde with visible dark roots",
      "style": "Long, straight, sleek, center part, draped over shoulders"
    },
    "face": {
      "makeup": "Full glam, sharp winged eyeliner, sculpted contour, nude glossy lips, defined eyebrows",
      "eyes": "Light colored (blue/green)"
    },
    "nails": "Long, almond-shaped, nude beige polish"
  },
  "outfit": {
    "top": {
      "type": "Strapless scarf top (handkerchief style)",
      "pattern": "{argument name="top pattern" default="Dior Oblique motif"} (navy blue and beige logo print)",
      "fit": "Fitted, cropped, showing midriff"
    },
    "bottoms": {
      "type": "Blue denim jeans",
      "style": "Low-rise, visible waistband"
    }
  },
  "accessories": {
    "jewelry": [
      "Small gold necklace with a horseshoe pendant",
      "Gold bangle bracelets on both wrists (stacked)",
      "Gold ring on right ring finger",
      "Silver belly button stud piercing"
    ],
    "bag": {
      "type": "Shoulder bag with gold chain",
      "material": "{argument name="bag material" default="Louis Vuitton Monogram canvas"} (brown)",
      "position": "Slung over right shoulder"
    },
    "tech": {
      "item": "iPhone Pro (triple camera lens)",
      "case_color": "Grey/Greenish solid color"
    }
  },
  "environment": {
    "setting": "Indoor, likely a bathroom or bedroom",
    "background": "Clean white walls, white door frame, minimalist aesthetic",
    "lighting": "Bright, cool-toned natural light, even illumination"
  }
}

Candid Mirror Selfie with Sequined Bikini and Anatomical Constraints

profile-avatar

A highly specific photorealistic prompt for generating a candid mirror selfie of a woman in a pink sequined bikini. It includes detailed descriptions of the subject's appearance, pose, and environment, with explicit negative prompts and constraints to maintain exact body volume and proportions, particularly focusing on the bust size.

展开提示词全文
{
  "subject": {
"description": "A woman with tanned skin and a dark, messy updo, holding a black smartphone. She has full lips with dark pink lipstick and dark eyes. Her skin has a noticeable sheen from oil. She wears a pink string bikini top with prominent, large pink sequined rose appliques and green leaf embroidery, and matching pink string bikini bottoms. Her body is toned with visible abdominal muscles. The chest volume is full and large, appearing natural and heavy.",
"physical_features": {
"hair": "Dark, messy updo, face-framing tendrils",
"skin": "Tanned, oiled sheen, realistic texture",
"face": "Full lips, dark eyes, confident expression",
"physique": "Toned, visible abs, prominent bust"
},
"clothing": {
"bikini_top": "Pink string, large sequined rose and leaf appliques",
"bikini_bottoms": "Pink string"
},
"accessories": [
"Black smartphone"
]
  },
  "pose": "A candid mirror selfie pose. She stands facing the camera, holding the smartphone up with her right hand, which is partially visible. Her left hand is placed firmly on her left hip. Her body is slightly angled, and her head is tilted slightly to the side.",
  "environment": "A modern indoor setting, likely a home. Behind her is a staircase with dark wood treads, white risers, and a metal and glass railing. The floor is dark wood. A window is visible in the background.",
  "lighting": "Cool, diffused indoor lighting. The light source appears to be coming from the window and overhead, creating distinct highlights on her oiled skin and catching the sequins on the bikini.",
  "mood_and_expression": "Confident, self-assured, candid. Her expression is direct and relaxed.",
  "style_and_realism": "Photorealistic, candid mirror selfie. The focus is sharp on the subject, with a slightly shallower depth of field in the background. No stylization.",
  "colors_and_tone": "Natural color palette. Dominant colors include the pink of the bikini, the tanned skin tone, the dark wood, and cool grey/white tones of the background.",
  "quality_and_technical_details": "High resolution, realistic texture fidelity, natural grain.",
  "aspect_ratio_and_output": "3:4",
  "controlnet": {
"pose_control": "enabled",
"depth_control": "enabled"
  },
  "negative_prompt": [
"anatomy normalization",
"dataset-average body proportions",
"reduced chest volume",
"flattened or compressed breasts",
"excessive slimming",
"beauty filters",
"plastic skin",
"professional studio lighting",
"over-stylized realism"
  ]
}
Maintain exact body volume and proportions from the reference image. Do not reduce chest size.

Style Transfer to Realistic Selfie

profile-avatar

A short instruction for users of Nano Banana Pro who want to perform a style transfer to a photorealistic aesthetic, specifically for images that start as a posed selfie.

展开提示词全文
Realistic POSED DYNAMIC SELFIE Photo Image

Candid Couple Portrait with Identity Lock

profile-avatar

A detailed image generation prompt for a candid, high-realism couple portrait set in a cozy indoor room. The prompt requires strict identity locking for both subjects based on uploaded reference images, specifying their clothing (white knit sweater, dark t-shirt) and a medium close-up framing with warm, soft indoor lighting.

展开提示词全文
{
  "image_prompt": {
    "reference": {
      "female_face": {
        "source": "use uploaded female reference image",
        "identity_lock": true,
        "face_accuracy": "exact match to reference — same facial structure, proportions, skin texture, makeup, and expression"
      },
      "male_face": {
        "source": "use uploaded male reference image",
        "identity_lock": true,
        "face_accuracy": "exact match to reference — same facial structure, beard, skin texture, and expression"
      }
    },

    "subjects": {
      "female": {
        "appearance": {
          "hair": "same length, color, and texture as reference",
          "makeup": "same makeup as reference",
          "skin": "same tone and natural texture"
        },
        "clothing": {
          "top": "white chunky knit sweater",
          "fit": "loose, cozy"
        },
        "pose": {
          "position": "standing slightly behind the man",
          "interaction": "arms wrapped gently around his shoulders",
          "expression": "soft, relaxed smile"
        }
      },

      "male": {
        "appearance": {
          "hair": "same as reference",
          "facial_hair": "same beard/moustache as reference",
          "skin": "same tone and texture"
        },
        "clothing": {
          "top": "dark brown or black t-shirt",
          "accessories": "cap worn forward (same style as reference)"
        },
        "pose": {
          "position": "standing in front",
          "interaction": "hands relaxed, natural posture",
          "expression": "calm, subtle smile"
        }
      }
    },

    "environment": {
      "location": "cozy indoor room",
      "background_elements": [
        "neutral walls",
        "minimal decor",
        "soft ambient lighting"
      ],
      "constraints": {
        "no_mirror": true,
        "no_reflections": true
      }
    },

    "camera": {
      "shot_type": "medium close-up",
      "angle": "eye-level",
      "framing": "both subjects centered and clearly visible",
      "lens": "natural perspective"
    },

    "lighting": {
      "type": "soft indoor lighting",
      "quality": "warm and even",
      "shadows": "gentle and flattering"
    },

    "style": {
      "realism": "high",
      "editing": "minimal",
      "aesthetic": "natural, candid couple portrait"
    },

    "constraints": {
      "no_face_change": true,
      "no_identity_merge": true,
      "no_beauty_overediting": true,
      "maintain_original_features": true
    }
  }
}

High-Contrast Monochrome Studio Portrait Prompt

profile-avatar

A prompt for generating a high-contrast, black-and-white studio portrait with cinematic film grain. It focuses on harsh, vibrant key lighting to sculpt facial features and highlight the texture of a raw denim shirt, emphasizing a relaxed, editorial mood and ultra-detailed photorealism.

展开提示词全文
Create a 1:1 high-contrast monochrome studio portrait with a half-body composition and cinematic film grain. A person with dark, swept-back hair makes direct eye contact, wearing a bright, genuine Duchenne smile. Harsh, vibrant key lighting sculpts sharp facial highlights and reveals the deep texture of a raw denim shirt with rolled sleeves. Underneath is a crisp white V-neck, paired with fitted chinos. Hands rest casually in the pockets. Set against a high-key white backdrop, the mood is relaxed and editorial, with ultra-detailed, photorealistic realism.

Desi Funny New Year 2026 Cinematic Portrait Prompt

profile-avatar

A prompt for generating a cinematic, ultra-realistic portrait with a 'desi funny' attitude for New Year 2026. It requires strict identity preservation from an uploaded image and instructs the AI to create a playful pose and outfit blending desi and modern styles, set against a festive night background with warm bokeh and subtle lighthearted elements.

展开提示词全文
{
  "prompt": "Use the uploaded image strictly as the identity reference. Preserve exact face match, facial structure, expressions, skin tone, hairstyle, and body proportions with no alteration. Create a desi-funny New Year 2026 cinematic portrait with playful swag. The subject poses in a humorous desi attitude—slight head tilt, cheeky grin, eyebrow raise, or playful hand gesture ({argument name="hand gesture" default="thumbs-up, peace sign, or pointing"}). Outfit blends desi and modern style: kurta-jacket, printed shirt, casual blazer, or festive wear with fun flair. Scene is set at night with a celebratory vibe—minimal background, warm lights, subtle bokeh, or festive glow. Add lighthearted elements like confetti, sparkler glow, or floating light orbs hinting at 2026 (no bold text). Lighting is cinematic but warm and friendly: soft key light on face, gentle fill, controlled highlights for a joyful look. Mood is funny, desi, confident, and celebration-ready. Ultra-realistic photography style, premium color grading, sharp focus, high dynamic range. Render in ultra high resolution, reel, DP, and poster-ready.",
  
  "negative_prompt": "face mismatch, identity change, extra people, cartoon, anime, illustration, clown style, blur, noise, harsh lighting, extra fingers, extra limbs, distorted anatomy, watermark, logo, text overlay",
  
  "style": "desi funny cinematic photography",
  "mood": "playful, desi swag, funny New Year 2026",
  "lighting": "warm cinematic lighting with soft glow",
  "background": "minimal festive backdrop with warm bokeh",
  "render_quality": "ultra high resolution, festive editorial quality",
  "composition": "centered subject with playful framing",
  "reference_strength": 0.9
}

Persian Carpet Inspired Portrait Prompt

profile-avatar

A prompt for generating an ultra-realistic portrait of a young girl, framed from the chest up, suitable for social media. The image features the subject wearing a dress inspired by Persian carpet patterns and holding a Persian cat, set against a luxurious Persian red carpet background with dramatic side lighting.

展开提示词全文
A 4K ultra-realistic portrait of a young girl, framed from the chest up (not full body), designed for Instagram story or profile picture. The background is a luxurious {argument name="carpet design" default="Persian red carpet design"} (only behind her, not on the ground). Sunlight shines from one side,casting natural warm light across half of her face. She is wearing an elegant dress inspired by Persian carpet patterns,with authentic Iranian colors and motifs. Her facial features remain unchanged.She is holding a {argument name="animal" default="Persian cat"} in her hands

Photorealistic Anatomy-Locked Mirror Selfie Prompt

profile-avatar

An extremely detailed JSON prompt using 'PX-LOCK-X' engine metadata to ensure zero drift in pose, anatomy, environment, and wardrobe when generating a photorealistic mirror selfie. The prompt specifies precise anatomical measurements and features of a highly muscular Afro-Caribbean male subject, aiming for a hyper-accurate reconstruction.

展开提示词全文
{
  "metadata": {
    "engine": "PX-LOCK-X",
    "prompt_version": "1.0.0",
    "generation_goal": "Photorealistic mirror-selfie recreation with zero drift across pose, anatomy, environment, lighting, and wardrobe",
    "determinism": {
      "pose_lock": true,
      "camera_lock": true,
      "composition_lock": true,
      "anatomy_lock": true,
      "environment_lock": true,
      "lighting_lock": true,
      "wardrobe_lock": true,
      "texture_lock": true,
      "deviation_tolerance": {
        "keypoint_px_at_4k": 6,
        "silhouette_iou_min": 0.94,
        "camera_shift_max_norm": 0.006
      }
    },
    "output_quality": {
      "resolution": "4K",
      "focus": "tack-sharp",
      "noise": "minimal"
    }
  },
  "subject": {
    "identity": {
      "gender": "male",
      "age_range": "early 30s",
      "ethnicity": "Afro-Caribbean",
      "height_estimate_ft_in": "6'1\""
    },
    "face": {
      "shape": "rectangular",
      "jawline": "broad, powerful",
      "cheekbones": "moderately high",
      "eyes": {
        "color": "dark brown",
        "gaze": "direct at phone",
        "openness": "relaxed, half-open",
        "expression": "calm, confident"
      },
      "nose": "straight with wide bridge",
      "lips": {
        "shape": "medium-full",
        "parting": "slightly parted"
      },
      "hair": {
        "style": "fully shaved head",
        "scalp_finish": "clean razor shave",
        "detail_lock": true
      },
      "facial_hair": {
        "type": "full beard",
        "length": "short boxed",
        "density": "thick",
        "grooming": "clean lines along cheeks and neckline"
      },
      "accessories": "none"
    },
    "skin": {
      "tone": "deep warm brown",
      "texture": "natural with subtle post-shower sheen",
      "body_hair": {
        "chest": "light, sparse",
        "abs": "minimal"
      },
      "moisture": "slightly damp"
    },
    "physique": {
      "build": "highly muscular athletic-bodybuilder hybrid",
      "proportions_inches": {
        "chest": 50,
        "waist": 33,
        "hips": 44
      },
      "upper_body": {
        "shoulders": "extremely broad",
        "neck": "thick and powerful",
        "pectorals": "dense, rounded",
        "deltoids": "deeply striated",
        "arms": "large with prominent vascularity",
        "triceps": "horseshoe-shaped definition"
      },
      "core": {
        "abs": "thick six-pack with deep separations",
        "obliques": "clearly carved",
        "adonis_belt": "strongly visible"
      },
      "lower_body": {
        "quadriceps": "thick and defined beneath towel",
        "fat_level": "very low"
      },
      "advanced_anatomy": {
        "thoracic_depth_in": 9.5,
        "waist_depth_in": 5.2,
        "hip_depth_in": 7.5
      },
      "symmetry": "high bilateral balance in muscle mass and vascularity"
    }
  },
  "pose": {
    "stance": "standing frontally",
    "orientation": "square to mirr

Sunkissed Glamour Tight Crop Selfie at Golden Hour

profile-avatar

A highly detailed, structured prompt for generating a hyper-photorealistic, intimate tight crop selfie with a 'Sunkissed Glamour' aesthetic. It simulates a professional mirrorless camera shot (f/1.4) with dramatic backlighting (contre-jour) and intense golden hour glow, focusing on extreme textural fidelity and a playful expression.

展开提示词全文
{
  "image_type": "Digital Self-Portrait",
  "shot": "Tight Crop Selfie",
  "shot_details": "Intimate, high-resolution lifestyle editorial portrait simulating a social media upload.",
  "style": "Sunkissed Glamour Aesthetic",
  "quality": "Hyper-photorealistic, Cinematic Quality, Extreme Textural Fidelity",
  "color_grade": "Warm, Golden Hour Tones, High Contrast due to backlighting, Silvery/White Highlights",
  "meta": {
    "aspect_ratio": "4:5",
    "resolution": "16K UHD"
  },
  "camera": {
    "device": "Professional Mirrorless Camera (Simulated)",
    "lens": "Wide-angle lens (28mm to 35mm equivalent)",
    "aperture": "f/1.4 (Extremely Shallow Depth of Field)",
    "distance": "Selfie range (Very close)",
    "angle": "Slight Low Angle",
    "framing": "Headshot/Bust Shot",
    "pov": "Direct First-Person Perspective",
    "focus": "Critical focus on the blue eyes",
    "lens_effect": "Diffuse Warm Lens Flares, Strong Backlight Glow, Subtle Chromatic Aberration"
  },
  "lighting": {
    "description": "Dramatic High-Key Contre-jour (Strong Backlighting)",
    "type": "Directional Sunlight",
    "source": "Intense sunlight streaming through large window",
    "primary": "Backlight creating severe contrast and silhouette effect on the face.",
    "secondary": "Soft, reflected fill light bouncing from bright white walls/environment.",
    "highlights": "Brilliant, silvery Rim Light halo along the contours of the hair and shoulder; fully Blown-out Highlights in the background (pure white).",
    "shadows": "Softly modeled shadows on the face, created by ambient fill light."
  },
  "scene": {
    "location": "Modern, minimalist interior",
    "environment": "Bright room with large, white window frames visible in the heavily overexposed background.",
    "time": "Daytime (simulating Golden Hour or intense midday sun)",
    "atmosphere": "Warm, luminous, airy, summer aesthetic."
  },
  "subject": {
    "gender": "Female",
    "age": "Early 20s",
    "ethnicity": "Nordic appearance",
    "appearance": "Tanned, healthy complexion",
    "body": {
      "skin": "Visible fine pores, sun-kissed tan, smooth texture."
    },
    "expression": {
      "eyes": "Clear light blue",
      "gaze": "Intense, direct eye contact with the camera",
      "mouth": "Slightly opened, wet glossy lips, tip of tongue playfully touching upper incisors.",
      "face_vibe": "Coquettish, playful, intimate."
    },
    "hair": {
      "color": "{argument name=\"hair color\" default=\"Long, platinum blonde\"}",
      "style": "Voluminous modern cut",
      "details": "Distinct, brighter face-framing strands."
    },
    "pose": {
      "description": "Frontal, intimate selfie pose, face directed slightly toward the light source.",
      "stance": "Close up, head and shoulder visible.",
      "energy": "Confident and inviting."
    }
  },
  "wardrobe": "Deep navy/black ribbed tank"
上一页 136 / 176 下一页

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 2026-10-28

内容有更新(细节请以官方文档为准)

内容更新 2026-10-27

内容有更新(细节请以官方文档为准)

内容更新 2026-10-26

内容有更新(细节请以官方文档为准)

内容更新 2026-10-24

内容有更新(细节请以官方文档为准)

内容更新 2026-10-23

内容有更新(细节请以官方文档为准)

内容更新 2026-10-22

内容有更新(细节请以官方文档为准)

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。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。