AI Image Prompts · 绘画提示词库

Create a 3D cartoon-style woman with a rap style (the face is inspired by the reference image). She should have a large head and oversized sneakers, an extra-large hoodie with a rap-style print, loose black pants, big sunglasses, and a thick chain necklace. Place her in a fun pose, with a white background, cinematic lighting, and a highly detailed Pixar-style render.

{
  "image_generation_prompt": {
    "subject_settings": {
      "placeholder": "{argument name="subject" default="[SUBJECT GOES HERE]"}",
      "positioning": "absolute dead-center composition, full subject visibility, no clipping, no cutoff edges",
      "presentation": "floating mid-air, isolated, no pedestals, no stands, no display boxes, no mannequins"
    },
    "aesthetic_style": {
      "visual_theme": "dreamy ethereal surrealism, vintage luxury editorial, 1990s high-fashion film photography",
      "texture": "heavy analog film grain, noisy texture, slight motion blur, soft focus dream-haze, chromatic aberration",
      "lighting": "starburst lens flare on metallic highlights, soft glowing aura, backlit silhouette depth, no studio lights, no softboxes"
    },
    "color_palette": {
      "primary_gradient": {
        "top_color": "#1A3A5F",
        "middle_color": "#7B9BB9",
        "bottom_color": "#D4823B"
      },
      "description": "A seamless dusk-to-dawn transition from deep midnight blue to a warm, hazy amber sunset glow"
    },
    "negative_constraints": [
      "text",
      "watermarks",
      "studio equipment",
      "lighting stands",
      "reflectors",
      "frames",
      "borders",
      "cut off subject",
      "out of frame",
      "multiple subjects",
      "cluttered background",
      "plastic textures",
      "hyper-realistic digital sharpness",
      "clean 4k render"
    ],
    "technical_parameters": {
      "aspect_ratio": "4:5",
      "style_strength": "high",
      "diffusion_iteration": "film_emulsion_simulation"
    }
  }
}

Viking
Kotogemoto
Mizukimoto

End of the Ministry of Magic

Eiji Voldemort
Mizuki Sirius Black

Land Cruiser, Ran Kuru, Ran Ku~ru

Harry Potter style
Slander

Viking
Completely in English (I didn't ask for that, lol)

Eiji Voldemort
Mizuki Sirius Black

'In the style of XX' is important, lol

Masterpiece, best quality, (3D cartoon render:1.5), (chibi mascot character:1.5), cute creature design, light novel cover composition.

(Main Subject: Dinosaur Transformation):
Center frame. The character is transformed into a cute, chubby dinosaur mascot (like a cartoon T-Rex or Yoshi).
Body features: Round body, short stubby limbs, thick tail, smooth scales texture. Big sparkling anime eyes (matching input color).
Costume: Wearing a wig matching input hair and ONLY the top/shirt/jacket from the input outfit. (No pants:1.5).

(Floating Spirit & Halo):
Floating beside the dinosaur's head is a (semi-transparent, ghostly spirit form:1.4) of The EXACT human character's face from the input image.
A glowing golden angel halo floats above this transparent head.
The expression on the spirit face is slightly bewildered or comical.

(Title Logo - Narou-Style):
At the top of the frame, a large, prominent, stylized JAPANESE title logo reads: "{argument name="Title Logo" default="I was reincarnated as a dinosaur in another world"}".
Typography style: Fantasy light novel design, bold outlines, magical flourishes, long title format, colorful.

(Style & Environment):
Blender 3D style, clay render texture, soft studio lighting, pastel colors.
Background: Simple, clean fantasy world element (e.g., stylized floating island or magic circle) but blurred to keep focus on subjects.

(Technical Tags): (3D render:1.4), (mascot:1.4), (chibi:1.3), (cute:1.4), (typography design:1.4), (ghost:1.3), (halo:1.3), sharp focus.
(Negative prompt): (human body on dinosaur:1.6), (human legs on dinosaur:1.5), (realistic dinosaur:1.5), (scary), (monster), (opaque spirit), (missing text), (missing halo), low quality, blurry, bad anatomy.

Viking
Kotogemoto
Mizukimoto

Renamed
Eiji Voldemort
Mizuki Sirius Black

Core Requirement: Convert all elements in the reference image—including characters, scenes, props, and environment—into ultra-realistic, live-action settings, resembling actual on-set movie stills.
Character Conversion: Strictly retain the original character's hairstyle, clothing, expression, facial features, body movements, and overall temperament, converting them into a real actor's image.
Scene Conversion:
• Restore the comic/illustration scene to a real, physically shootable space
• All props, furniture, and architecture must be real set construction or real location shooting
• Realistic material texture: fabric folds, wood grain texture, metal reflections, wall texture
• Rich environmental details: dust, signs of wear, everyday life details
• Spatial perspective and light/shadow must conform to physical laws
Photography Style: Cinematic live-action photography
• Deep depth of field composition, with everything clear from foreground to background
• Film texture, soft grain, and rich color layering
• Dramatic light and shadow contrast: warm tungsten working lights vs. cool natural twilight
• Subject prominence while retaining environmental narrative
• Scene art and photography comparable to 'Blade Runner' or 'The Grand Budapest Hotel'
Technical Specifications:
• Film Stock: Kodak Vision3 500T (cinematic photography standard)
• Applicable Scenes: Night scenes, low light environments, indoor tungsten lighting
• Aperture: f/8–f/11 (deep depth of field, entire frame clear)
• Focal Length: 35mm–50mm (standard cinematic perspective)
• Resolution: 4K cinema or higher
• Color: Warm, saturated tungsten color temperature, retaining detail in shadows
• Lighting Design: Three-point lighting or natural light combined to create cinematic light and shadow layers
Scene Art Requirements:
• Real set construction or location shooting, avoiding the feel of CG synthesis
• Prop placement conforms to the character's background and era setting
• Spatial design with depth and a sense of life
• Environmental elements like walls, floors, and ceilings must have realistic texture
• Background characters (if any) must also be real actors, avoiding blurring
Negative Prompts (Strictly Avoid):
• Blurred background, shallow depth of field, bokeh, out-of-focus
• Cartoon, anime, illustration style, hand-drawn texture, CG rendering feel
• Flat composition, lack of spatial depth
• Fake sets, green screen artifacts, digital backgrounds
• Facial deformation, disproportionate scaling, AI generation artifacts
• Low quality, digital noise, pixelation, excessive sharpening
• Overexposed strong light, plastic texture, excessive digital post-processing
• Floating props, physical implausibility, material texture mapping feel
• Empty background, lack of detail, monotonous environment
Reference Style: Masterpieces of cinematic photography - the live-action aesthetic of Roger Deakins, Emmanuel Lubezki, Hoyte van Hoytema.

A whimsical, close-up high-detail scene featuring a classic pair of glossy red, wind-up chattering teeth with perfectly white, reflective teeth, the jaws wide open as if mid-sentence, captured in sharp focus. A polished silver winding key protrudes from the side of the upper jaw, catching soft, warm light. Attached beneath the teeth is a pair of cartoonish, slightly oversized tan feet with smooth skin and bright red, glossy polish on all ten toes, posed as if taking a playful step forward, suggesting motion. The chattering teeth stand inside an overflowing toy box made of painted wood, its interior cluttered with colorful toys—tiny cars, sparkling plastic jewelry, wrapped candies, rubber balls, and scattered trinkets—some spilling over the edge. Soft, diffused lighting creates gentle shadows and a cozy, nostalgic atmosphere, with a shallow depth of field subtly blurring the background toys to keep the focus on the animated chattering teeth and feet.

highly detailed pebble mosaic portrait of a beautiful woman, where the entire image is constructed from thousands of smooth natural stones and pebbles. The pebbles vary in size and color to form realistic facial features, hair, skin tones, and clothing patterns. The woman has long flowing hair and is wearing a richly patterned red traditional dress. The background is also composed of carefully arranged river stones creating a textured mosaic wall. Ultra-detailed stone texture, natural earthy colors, soft lighting, artistic stone mosaic craftsmanship, gemini photorealistic composition, 8k resolution, intricate pebble placement, handcrafted mosaic art style.

an anthropomorphic pangolin, standing upright in a well-fitted tailored suit and tie, holding a neatly arranged bouquet of fresh flowers, captured mid-motion as he reaches out to ring the ornate doorbell of a cozy suburban house, early evening soft golden light, warm welcoming atmosphere, detailed scales and fabric textures, cinematic composition

Ultra wide-angle lens cinematic photo captured from the perspective of a rabbit that is about to be caught in the talons of a Golden Eagle

UNIVERSAL MONO PUNK SCENE GENERATOR

VARIABLE GUIDE

1. PUNK_STYLE =
   [Choose one punk aesthetic. If left blank, one style will be automatically selected at random from LIST_OF_PUNK_STYLES.]

2. LIST_OF_PUNK_STYLES =
   [Steampunk, Dieselpunk, Atompunk, Cyberpunk, Biopunk, Nanopunk, Solarpunk, Decopunk, Clockpunk, Coalpunk, Raypunk, Stonepunk, Sailpunk, Cassettepunk, Teslapunk, Mythpunk, Sandalpunk, Oceanpunk, Frostpunk, Junglepunk, Desertpunk, Gothicpunk, Ironpunk, Crystalpunk, Rustpunk, Vaporwavepunk, Neonpunk, Skypunk, Aquapunk, Arborpunk, Astropunk, QuantumPunk, Mechanopunk, Technopunk, Retrocomputingpunk]

3. SUBSTYLE =
   • If a specific SUBSTYLE is provided → use it exactly as written.
   • If SUBSTYLE is set to **none** → do not apply a substyle layer.
   • If SUBSTYLE is left **blank** → automatically infer a logical SUBSTYLE based on the chosen PUNK_STYLE.

4. TECHNOLOGY_LEVEL =
   • If a specific TECHNOLOGY_LEVEL is provided → use it exactly as written.
   • If TECHNOLOGY_LEVEL is set to **none** → do not apply a technology modifier.
   • If TECHNOLOGY_LEVEL is left **blank** → automatically infer an appropriate technology level based on the chosen PUNK_STYLE.

5. ENVIRONMENT =
   • If a specific ENVIRONMENT is provided → use it exactly as written.
   • If ENVIRONMENT is set to **none** → do not apply an environmental modifier.
   • If ENVIRONMENT is left **blank** → automatically infer a logical environment based on the other active layers.

6. CULTURE =
   • If a specific CULTURE is provided → use it exactly as written.
   • If CULTURE is set to **none** → do not apply a cultural modifier.
   • If CULTURE is left **blank** → automatically infer a cultural framework appropriate to the world.

7. ROOM_OR_SETTING =
   [IF BLANK: Generate a compelling location consistent with the active world layers.]

8. LIST_OF_OBJECTS =
   [IF BLANK: Infer 4–6 logical objects belonging to the environment.]

9. CAMERA_FRAMEWORK =
   [IF BLANK: Infer the best cinematic framing for the environment.]

10. LIGHTING_FRAMEWORK =
    [IF BLANK: Infer lighting derived from the technology and environment.]

11. MATERIAL_SYSTEM =
    [IF BLANK: Infer dominant materials and structural systems derived from the style layers.]

---

### STYLE SELECTION DIRECTIVE
If **PUNK_STYLE is left blank**, randomly select **one style from LIST_OF_PUNK_STYLES** and apply it as the active world aesthetic.
All subsequent inference systems must use the chosen PUNK_STYLE.

### WORLD INTERPRETER ENGINE
Analyze all active layers:
PUNK_STYLE
* optional SUBSTYLE
* optional TECHNOLOGY_LEVEL
* optional ENVIRONMENT
* optional CULTURE

From these layers infer:
• dominant materials
• architectural forms
• infrastructure systems
• lighting technology
• environmental atmosphere
• cultural design motifs

All visual elements must remain internally consistent with this world model.

### SCENE ENGINE
Generate a highly detailed, cinematic, p