Original document
GPT Image 2 · en
Ultra-realistic cozy Japanese-Korean café photography featuring a cute young [Japanese/Korean] couple sitting together naturally in a trendy aesthetic café. The young couple should look stylish and youthful, wearing [fashion style/outfit colors], smiling softly and enjoying desserts together.
The table is beautifully filled with [desserts/foods] such as pancakes, strawberry cakes, macarons, croissants, pastries, iced coffees, matcha lattes, fruit desserts, and aesthetic drinks arranged in a visually satisfying composition. Add small aesthetic café props like [flowers/ribbons/books/candles/pearls/notebooks] on the table for a premium Pinterest moodboard feel.
Soft [lighting style] lighting enters through the café windows creating dreamy highlights, creamy shadows, glossy reflections on drinks, and realistic dessert textures. Background should contain softly blurred [Japanese/Korean] signs, glowing café boards, handwritten Japanese text, neon typography, and aesthetic city café elements for an authentic Tokyo/Seoul vibe.
Add cute scrapbook-style doodles and handwritten notes around the image in [doodle color] ink — tiny hearts, stars, sparkles, ribbons, arrows, smiley sketches, bows, diary stickers, and handwritten café notes.
Color palette should focus on [color theme] tones. Style inspired by viral Pinterest café photography, Korean lifestyle aesthetics, Japanese cozy café culture, dreamy Gen-Z romance mood, shallow depth of field, cinematic composition, ultra realistic food textures, soft blurry background, ultra detailed realistic photography, clean aesthetic layout, 8k.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.