
Rainy Japanese Street in a Snow Globe
A moody neon-lit Japanese alley shrunk to snow globe scale, nostalgic, cinematic, and endlessly scrollable.
The prompt
Create a hyper-realistic 3D miniature diorama that looks like a collectible snow globe or scale model kit. The scene shows a narrow rainy Japanese alley at night, a tiny ramen shop glows with warm orange neon kanji signs, paper lanterns hang between buildings, miniature puddles reflect the neon light with perfect tiny mirror-like clarity, a lone figure with a transparent umbrella stands at the entrance. Rain streaks are visible as fine silver lines. The entire world fits inside what appears to be a 10cm diorama base with visible wooden edges and a brass nameplate. Tilt-shift bokeh blurs the background buildings softly. Style: photographic, studio-lit product shot of the diorama on a dark slate surface, 4K detail.
Why it goes viral
Neon reflections in miniature puddles + the visible diorama base trick the eye into craving this as a real product to buy.
How to use it
Paste the full paragraph prompt into ChatGPT with GPT-4o image generation enabled. If the diorama base is missing, add: 'Show the wooden diorama base and brass label clearly at the bottom.' Request a second variation with a slightly wider shot to show more of the alley depth. Crop to 4:5 in-app and post with the caption asking followers to name the ramen shop.
Pro tips
Describe the physical base ('wooden edges', 'brass nameplate'), GPT-4o treats the diorama as a real object and renders its housing. Specifying the diorama size ('fits in 10cm') anchors the scale and prevents the scene from looking like a full-size street. Ask for a 'studio-lit product shot on dark slate' as the framing, this immediately lifts the image to editorial quality. Add 'fine silver rain streaks' rather than just 'rain' for crisp, visible precipitation at miniature scale. If the first output lacks neon reflections in puddles, follow up with: 'Regenerate keeping the same scene but make the puddle reflections sharper and more mirror-like.'
Steps and settings
- 01Paste the full paragraph prompt into ChatGPT with GPT-4o image generation enabled.
- 02If the diorama base is missing, add: 'Show the wooden diorama base and brass label clearly at the bottom.'
- 03Request a second variation with a slightly wider shot to show more of the alley depth.
- 04Crop to 4:5 in-app and post with the caption asking followers to name the ramen shop.
- 3D & Toys
- neon reflections
- Japanese alley
- rain
- snow globe aesthetic
- tilt-shift bokeh
- product shot
- night scene



