How to Make a Studio Ghibli Photo with AI — Complete 2026 Guide

Studio Ghibli photos have taken over social media — and for good reason. The soft watercolor textures, warm golden light, and hand-painted aesthetics of Miyazaki's films translate beautifully into AI-generated images. This guide walks you through exactly how to create Ghibli-style photos using ChatGPT 4o and Midjourney v7, including 10 copy-paste prompts and tips to get consistent results.

What Makes Studio Ghibli Style Unique?

Studio Ghibli films are defined by a specific visual language that took Hayao Miyazaki decades to perfect. When recreating this style with AI, you need to communicate these visual elements precisely through your prompt.

• Soft, hand-painted linework with visible brushstroke texture
• Watercolor-washed backgrounds with soft color bleeding at edges
• Warm, muted color palette — sage greens, dusty blues, golden yellows
• Organic, flowing shapes — no harsh edges or sharp geometry
• Natural settings: forests, meadows, misty mountains, rural villages
• Expressive characters with simplified but emotionally rich faces
• Soft, diffused golden-hour or overcast lighting

The most common mistake beginners make is writing a short prompt like 'Ghibli style portrait.' The AI doesn't know which aspects of Ghibli you want — the color palette? The linework? The background style? A good Ghibli prompt specifies at least 4–5 of these elements explicitly.

Method 1: ChatGPT 4o (Best for Transforming Existing Photos)

ChatGPT 4o is the easiest tool for Ghibli transformations because you can upload your actual photo and ask it to transform it. The model preserves facial identity well, making it ideal for portraits and couple photos.

• Open ChatGPT 4o (free tier supports image generation)
• Click the paperclip icon to upload your photo
• Paste the prompt below — do not modify the structure, only change the subject details if needed
• If the first result doesn't match, add: 'More watercolor texture, less photorealistic'
• For group photos, add: 'Preserve all faces and identities exactly'

ChatGPT Ghibli Prompt — Portrait

Transform this photo into Studio Ghibli animation style.
Subject: portrait of a person.
Lighting: soft golden hour sunlight filtering through leaves.
Mood: serene, peaceful, and nostalgic atmosphere.
Environment: lush forest clearing with wildflowers and soft mist.
Camera: medium close-up, eye-level.
Additional qualities: soft hand-painted linework, watercolor-washed background, organic flowing shapes, warm muted color palette, Hayao Miyazaki aesthetic, visible brushstroke texture.
Preserve the subject's facial identity and key features exactly.

ChatGPT Ghibli Prompt — Couple

Transform this photo into Studio Ghibli animation style.
Subject: couple portrait.
Lighting: warm romantic golden hour light.
Mood: tender, romantic, and softly nostalgic.
Environment: hilltop meadow with tall grass and a distant village below.
Camera: wide medium shot.
Additional qualities: soft hand-painted linework, watercolor-washed background, warm muted palette, flowing organic shapes, Miyazaki aesthetic.
Preserve both subjects' facial identities and features exactly.

Method 2: Midjourney v7 (Best for Pure Ghibli Art)

Midjourney v7 produces the most stylistically consistent Ghibli results. It won't transform an uploaded photo the same way ChatGPT does, but for generating original Ghibli-style images from scratch, the quality is unmatched.

• Open Discord and go to a Midjourney channel
• Type /imagine and paste the prompt
• Add --ar 4:5 for portrait format, --ar 16:9 for landscape
• Add --style raw to reduce Midjourney's default processing
• Use --v 7 to ensure you're on the latest model

Midjourney Ghibli Prompt — Portrait

Studio Ghibli animation style portrait, soft hand-painted linework, watercolor-washed background, organic flowing shapes, warm muted color palette, golden hour sunlight, serene peaceful atmosphere, lush forest clearing, Hayao Miyazaki aesthetic, brushstroke texture --ar 4:5 --style raw --v 7

10 Ready-to-Use Ghibli Prompts

Copy any of these directly into ChatGPT or Midjourney. They're pre-tested for consistent results:

• Ghibli portrait — golden hour: 'Transform this photo into Studio Ghibli style. Soft golden light, forest clearing, serene mood, hand-painted linework, watercolor background. Preserve facial identity.'
• Ghibli couple — hilltop: 'Ghibli animation style couple. Hilltop at sunset, wildflowers, warm romantic light, Miyazaki aesthetic. Preserve both faces.'
• Ghibli child — playful: 'Ghibli style child portrait. Soft afternoon light, garden with giant sunflowers, joyful curious mood, watercolor texture.'
• Ghibli pet — cozy: 'Ghibli style pet portrait. Warm indoor light, cozy cottage setting, wooden furniture, soft illustrated look.'
• Ghibli landscape — misty mountains: 'Ghibli animation landscape. Misty mountain range, ancient forest, small village in valley, overcast light, painterly style.'
• Ghibli portrait — rainy day: 'Ghibli style rainy day portrait. Soft grey light, rain on window, cozy interior, nostalgic melancholic mood, hand-painted texture.'
• Ghibli product — food: 'Ghibli style food illustration. A steaming bowl of ramen, warm kitchen light, cozy atmosphere, illustrated painterly style.'
• Ghibli childhood memory: 'Ghibli animation style childhood scene. Summer afternoon, cicadas, old wooden house, a child looking at a distant horizon.'
• Ghibli portrait — night: 'Ghibli style night portrait. Soft moonlight, fireflies, dark forest, magical mystical mood, hand-painted linework.'
• Ghibli landscape — coastal: 'Ghibli animation coastal scene. Rocky cliffs, deep blue ocean, seagulls, morning mist, painterly watercolor style.'

Tips for Better Ghibli Results

• Always specify the environment — Ghibli style without a background setting looks generic
• Use mood words like 'serene', 'nostalgic', 'melancholic', 'whimsical' — they influence color temperature and composition
• For ChatGPT: if the result is too photorealistic, add 'more illustrated, less photorealistic, visible brushstrokes'
• For Midjourney: --style raw prevents over-processing and keeps the hand-painted look
• High-quality source photos (well-lit, sharp) produce significantly better transformations
• Avoid sunglasses or heavy accessories in source photos — they confuse the model

Common Mistakes to Avoid

• Writing 'Ghibli style' without any other details — too vague, results are inconsistent
• Using dark, heavily shadowed source photos — Ghibli style requires soft, readable lighting
• Forgetting to say 'preserve facial identity' for portrait transformations
• Using Midjourney for photo transformation (it doesn't work the same way as ChatGPT)
• Adding too many conflicting styles — 'Ghibli + cyberpunk + oil painting' confuses the model

Studio Ghibli prompts consistently produce some of the most beautiful AI art when done right. The key is being specific about the lighting, environment, and mood — three elements that define Ghibli's visual language more than anything else. Start with the prompts above and adjust from there.