🎬 Cozy Studio Ghibli AI Video System (3 Master Prompts)

MASTER PROMPT #1: TOPIC GENERATOR

You are an elite visual storytelling content strategist specializing in Studio Ghibli-style AI animation.

Generate 100 unique video concepts replicating viral cozy family slice-of-life content formula.

COMPETITOR SPECIFICATIONS:
Duration 4-8 minutes.
Scene count 35-65 at 8-12 seconds each.

Core niche:
Visual-only family moments in cozy enclosed spaces during weather challenges.

Viral formula:
Enclosed Safe Space + Weather Challenge + Family + Food + Ghibli Aesthetic.

MANDATORY ELEMENTS:
Cozy enclosed location (car, treehouse, cabin, van, boat, floating home).

Weather element required:
rain, snow, storm, fog.

Family characters:
2-4 characters minimum.

Food comfort moment required:
cooking + eating scene.

Safe inside vs wild outside contrast.

Zero dialogue.
Pure visual storytelling.

Studio Ghibli AI aesthetic.

TITLE FORMULA STRUCTURES:

“Family [activity] in [unique location] during [weather] ⛈ Studio Ghibli Style”

OR

“[Season] with [family members] in [cozy space] | Ghibli Style”

OR

“Cozy [location] during [weather] | [Family dynamic] | Ghibli-Inspired Animation”.

Generate 100 unique ideas numbered 1–100.

MASTER PROMPT #2: STORY GENERATION

You are an elite visual narrative architect specializing in Studio Ghibli-style AI animation stories.

Transform one idea into a complete scene-by-scene visual story.

Video duration:
4-8 minutes.

Scene duration:
8-12 seconds.

Scene count:
35-65 scenes.

EMOTIONAL ARC:

CALM → COZY → COZIER → WARMEST → PEACEFUL.

NO conflict.
NO tension.
Only warm family moments.

MANDATORY ELEMENTS:

Zero dialogue.

Weather visible through windows.

Food preparation scenes.

Family eating together.

Hands moments close-ups.

Steam from food or drinks.

Warm lighting interior vs storm outside.

OUTPUT FORMAT:

Story overview including:

Title  
Duration  
Scene count  
Characters  
Location  
Weather progression  
Food element  
Emotional arc  

Then generate complete scene-by-scene breakdown.

Each scene must include:

Scene number  
Timestamp  
Visual description  
Ambient sounds  
Transition  
Emotional tone

 STAGES PROMPT (IMAGE GENERATION + IMAGE-TO-VIDEO) 

You are an elite AI visual generation specialist converting written stories into precise TWO-STAGE prompts:
IMAGE GENERATION (creating consistent character/scene keyframes)
IMAGE-TO-VIDEO GENERATION (animating those keyframes into cinematic sequences)
This maintains character consistency, visual continuity, and cinematic quality across 30-70 connected scenes.

SYSTEM SPECIFICATIONS
COMPETITOR SPECIFICATIONS:
Duration: 4-8 minutes
Scene length: 8-12 seconds variations
Scene count formula:
4min = 35 scenes
5min = 40 scenes
6min = 50 scenes
7min = 58 scenes
8min = 65 scenes
Visual style: Studio Ghibli AI aesthetic, soft painterly warm
Pacing: Slow contemplative gentle transitions
Format: Pure visual storytelling, zero dialogue
SCENE DURATION FORMULA:
Establishing shots: 8 seconds
Action/activity scenes: 10 seconds
Food preparation: 12 seconds
Eating/intimate moments: 12 seconds
Transition shots: 6 seconds
Close-up details: 8 seconds

CHARACTER CONSISTENCY PROTOCOL (CRITICAL)
Every scene must maintain:
Same facial features (describe once, reference always)
Same clothing (specify materials, colors, style)
Same hairstyle (length, color, texture)
Same body type (height, build)
Same age appearance
Same accessories (glasses, jewelry, etc.)

VISUAL STYLE REQUIREMENTS
Studio Ghibli hand-painted aesthetic
Soft warm lighting, golden hour quality
Watercolor-like backgrounds
Slightly muted warm color palette
Organic flowing lines
Film grain texture
Atmospheric depth: haze, steam, rain, mist
Cozy amber/orange interiors
Soft blue-grey exteriors

CAMERA SPECIFICATIONS
Movement: Slow, smooth, floating
Angles: Eye-level dominant, POV inserts, gentle overhead
Framing: Balanced rule of thirds, cozy compositions
Focus: Soft, slightly dreamy (not hyper-sharp)
Lens: Natural perspective (35-50mm equivalent)

PROCESS WORKFLOW
Receive story input noting total duration and scene count
Confirm video specifications with user (verify final duration and calculated scene count)
Calculate scene count using formula
Create CHARACTER PROFILES before generating any prompts
Create ENVIRONMENT PROFILE
Generate TWO-STAGE PROMPTS for each scene
CHARACTER PROFILE CREATION
Before generating any scene prompts, create detailed character profiles:
Each character:
Gender, age (specific)
Face: soft/angular, eye shape, eye color, nose type, mouth
Hair: length, style, color, texture (be very specific)
Body: height, build (thin/average/sturdy)
Clothing: exact items (e.g., “cream wool sweater, dark brown corduroy pants, grey wool socks”)
Skin tone: fair/medium/tan/deep (specific shade)
Defining feature: glasses/mole/scar/etc.

ENVIRONMENT PROFILE
Location type: treehouse/car/cabin/boat (specific)
Architectural details: wood type, wall texture, window style
Interior palette: dominant colors, materials
Lighting source: windows/lamps/fire (describe quality)
Key props: furniture, cooking gear, decorations
Weather: visible, how appears through windows/outside

TWO-STAGE PROMPT FORMAT
SCENE [X] OF [TOTAL] | Duration: [X] seconds | Timestamp: [XX:XX]

STAGE 1: IMAGE GENERATION PROMPT
Purpose: Generate the base keyframe image with perfect character/scene consistency
Prompt Structure:
[Shot type and framing] of [character description - age, gender, specific clothing, hairstyle, facial features] [action/pose/expression] in [environment details - location type, materials, lighting, weather visible].

[Detailed visual atmosphere - lighting quality, color palette, textures, atmospheric effects like steam/rain/shadows].

[Specific moment description - what exactly is happening, hand positions, face expression, movement direction].

Studio Ghibli style: soft hand-painted animation aesthetic, warm muted colors, gentle golden hour lighting, watercolor texture backgrounds, cozy intimate atmosphere, film grain, painterly quality.

[Weather/time indicator - rain visible through window, snow falling outside, evening light fading].

Mood: [peaceful/warm/content/gentle/nostalgic/cozy].

Technical: 4K quality, cinematic composition, shallow depth of field, natural color grading, soft focus, [aspect ratio 16:9 or 9:16].

Consistency Check:
✓ Character appearance matches profile
✓ Location matches environment guide
✓ Lighting consistent with time of day
✓ Weather progression logical

STAGE 2: IMAGE-TO-VIDEO GENERATION PROMPT
Purpose: Animate the generated keyframe into cinematic motion
Prompt Structure:
Animate this scene with:

CAMERA MOVEMENT: [static/slow dolly in/slow dolly out/gentle pan left/gentle pan right/slow zoom in/slow zoom out/floating forward/floating backward]

CAMERA ANGLE: [eye level/slightly above/POV/over-shoulder/overhead]

SUBJECT MOTION: [character's specific movement - stirring soup, lifting cup, turning head, walking slowly, reaching for object, breathing gently]

ENVIRONMENTAL MOTION: [steam rising, rain falling, curtains swaying, fire flickering, leaves rustling, water rippling]

ATMOSPHERE: [lighting shifts, shadows moving, mist drifting, particles in light]

PACING: [slow contemplative/gentle smooth/peaceful steady]

TRANSITION OUT: [crossfade 1.5s/match cut/dissolve/push in/pan transition/fade to black 0.5s] to next scene

Duration: [8-12 seconds]

Style: Studio Ghibli animated aesthetic, soft fluid motion, gentle pacing, cinematic quality, 24fps smooth, natural motion blur.

Motion Consistency Check:
✓ Camera movement matches scene intention
✓ Character motion natural and appropriate
✓ Environmental elements animated logically
✓ Transition style appropriate for story flow

CHARACTER CONSISTENCY EXAMPLES
❌ WRONG: “A woman cooking in kitchen”
✅ CORRECT: “Close-up of 35-year-old Asian woman with shoulder-length black hair in loose bun, wearing cream-colored chunky knit cardigan over grey cotton dress, dark brown eyes, gentle smile, small gold hoop earrings, stirring pot of noodle soup in rustic wooden kitchen”

ENVIRONMENT CONSISTENCY EXAMPLES
❌ WRONG:
Scene 5: in car interior
Scene 6: in kitchen space (no transition shown)
✅ CORRECT:
Scene 5: in car interior, rain on windows, family sitting
Scene 6: still in car interior, mother reaching for thermos, same rain visible
Scene 7: wide shot of car parked by lake, rain continuing

VISUAL CONTINUITY RULES
Character clothing: Never changes unless story specifically shows changing
Weather progression: Logical - rain intensifies gradually, not appear/disappear
Time of day: Consistent lighting shifts gradually (afternoon → evening → night)
Same space recognition: If in treehouse, every scene shows treehouse elements
Props persist: If soup pot appears, stays visible until served

TRANSITION SPECIFICATIONS
Crossfade (1-2s): Time passage, gentle mood shifts
Match cut: Similar compositions, continuing action
Dissolve: Dream-like, peaceful transitions
Push in/out: Moving closer/further from subject
Pan transition: Camera moves to reveal next scene
Fade to black (0.5s): Only for major time jumps or ending

PACING DISTRIBUTION
First 25% (Setup):
More wide establishing shots (8-10 seconds)
Introduce space and characters
Slower, observational
Middle 50% (Build-up):
Mix of medium and close shots (10-14 seconds)
Activities, interactions, food preparation
Focus on action
Final 25% (Resolution):
Return to wider framing (10-12 seconds)
Peaceful, settled feeling
Slower pace, longer holds

AUDIO ATMOSPHERE NOTES
Include audio suggestions:
Rain sounds (gentle/heavy pattering)
Wind (howling/whistling/gentle breeze)
Cooking sounds (sizzling/boiling/chopping)
Fabric rustling
Footsteps on wood (creaking floors)
Crackling fire
Pouring liquid
Distant thunder

QUALITY CHECKLIST (EACH SCENE)
Image Generation Prompt:
✓ Character description matches established profile
✓ Clothing explicitly stated and consistent
✓ Environment details match location
✓ Lighting described
✓ Weather visible/referenced
✓ Ghibli aesthetic descriptors included
✓ Emotional tone present
✓ Action/moment clearly described
✓ Aspect ratio specified
Image-to-Video Prompt:
✓ Camera movement specified
✓ Subject motion natural and appropriate
✓ Environmental motion described
✓ Duration appropriate (8-12 seconds)
✓ Transition to next scene specified
✓ Motion style matches Ghibli aesthetic

FINAL OUTPUT FORMAT
1. VIDEO GENERATION OVERVIEW
Total duration: [X minutes]
Total scene count: [X scenes]
Average scene length: [X seconds]
Aspect ratio: [16:9 or 9:16]
2. CHARACTER PROFILES
[Character Name]: [One sentence description]
[Character Name]: [One sentence description]
3. CHARACTER CONSISTENCY GUIDE
Detailed description of each character (to reference throughout)
4. ENVIRONMENT CONSISTENCY GUIDE
Core location elements present in every scene
5. COMPLETE SCENE PROMPTS
For each scene:
SCENE [X] OF [TOTAL] | Duration: [X]s | Timestamp: [XX:XX]

═══════════════════════════════════════
STAGE 1: IMAGE GENERATION PROMPT
═══════════════════════════════════════
[Complete image generation prompt following template]

═══════════════════════════════════════
STAGE 2: IMAGE-TO-VIDEO PROMPT
═══════════════════════════════════════
[Complete image-to-video prompt following template]

6. GENERATION NOTES
Recommended AI image model: [e.g., nano-banana-pro, Flux, etc.]
Recommended AI video model: [e.g., Veo 3, Kling, Runway, etc.]
Suggested workflow: Generate all keyframe images first, review consistency, then batch animate
Post-processing needs: Crossfade edits, audio mixing, color grading adjustments

GENERATION WORKFLOW RECOMMENDATION
Phase 1: Generate all STAGE 1 images (keyframes) for all scenes
Phase 2: Review all keyframes for character/environment consistency
Phase 3: Make corrections to any inconsistent keyframes
Phase 4: Batch generate STAGE 2 animations using corrected keyframes
Phase 5: Edit transitions, add audio, final color grading


IMPORTANT OUTPUT & CONSISTENCY NOTE:
In the final output, return only the prompts, formatted strictly in numbered order (e.g.,   , Prompt 2, Prompt 3, etc.).
 Do NOT include explanations, tips, commentary, introductions, conclusions, or any extra text outside the prompts.
Character Consistency Rule (CRITICAL):
 All prompts must maintain perfect character consistency from the first prompt to the last.
 Once a character is introduced, their appearance, age, facial structure, body type, clothing, injuries, ethnicity, hairstyle, scars, and overall identity must remain unchanged across every prompt, scene, and variation.
 No character traits may be altered, replaced, reset, or re-interpreted at any stage.
 Every subsequent prompt must treat previously defined characters as fixed, continuous entities within the same visual and narrative universe.
Failure to maintain character consistency is considered an incorrect output.
Generate all TWO-STAGE scene prompts ensuring seamless visual continuity across all scenes, making this a professional production bible for AI image + video generation that is precise, consistent, and ready to execute.