Scene-to-diorama miniature
Turns an uploaded scene photo into a clean, collectible-scale miniature diorama that keeps the original layout and object placement.
Budget models do not render typography reliably; this case needs a higher tier.
Image generated by us with gpt-image-2. Prompt by @egeberkina
Prompt
Reconstruct the uploaded scene photograph as a clean, collectible-scale miniature diorama, preserving the exact layout, object placement, and spatial relationships from the reference image. Rebuild every vehicle, structure, prop, terrain feature, and figure visible in the photo at faithful miniature proportions, simplifying detail while keeping the composition unchanged. Mount the diorama on a [BASE_SHAPE] sculpted base with clean, minimal, slightly beveled edges, resting on a seamless [BACKDROP_COLOR] studio backdrop with no environment beyond the base. Frame the whole diorama in an isometric or slightly elevated three-quarter view, fully visible and centered, with no zoom and no cinematic cropping. Light the scene with soft, [LIGHTING_MOOD] studio lighting from overhead, angled slightly, producing only subtle contact shadows beneath the objects; avoid dramatic lighting, spotlights, or cinematic contrast. Render any figures as [FIGURINE_STYLE] miniature figurines with no facial realism. Use smooth, matte-to-satin surface finishes and high-quality scale-model materials for a clean, modern, premium diorama look. Keep the final image ultra-clean with no grain, noise, or motion blur, and do not include a glass case, display box, enclosure, reflections, labels, plaques, or text overlays.
4 placeholders still empty. Copied as is, the generator will draw the brackets literally inside the image.
Editable variables
Change any value and the prompt rewrites above. Copy the finished result.
The outline shape of the sculpted diorama base.
Colour of the seamless studio background behind the diorama.
Overall tone of the soft studio lighting.
Finish applied to any miniature figures rebuilt from the scene.
Generate the image here
With your own OpenAI key. Free for us, and what you spend is yours.
Kept in this browser only and sent straight to OpenAI's API: it never passes through a server of ours, because there is no server of ours in between. Check it yourself in the network tab.
Fill in the remaining variables before generating, or the model will draw the brackets.
