Generate Similar

Generation Mode Choose how the image should start.

Prompt

2x2 grid, do this for 4 famous Middle Eastern dishes: 
Create ONE square (1:1) image: premium product-photography of a “miniature life” culinary construction diorama inside a crystal-clear five-sided acrylic display cube on a light wood base. The scene must read as a real tabletop studio product shot.
 
Using only the dish name <Input A> and authentic culinary knowledge, infer:
4–5 canonical raw ingredients that are visually distinct (color/shape/texture).
1 “hero prop” container associated with making or serving this dish (a single large can/jar/bottle/box that makes sense for <Input A>).
1 central appliance/tool that best represents how the dish is cooked.
Enforce consistency:
The inferred raw ingredients must appear both in the ingredient cubbies and be visibly represented in the finished dish.
The hero prop must logically match one inferred ingredient or key component.
The appliance must match the cooking method for <InputA>.
Container  (the recipe box):
A crystal-clear acrylic display cube (five-sided) with subtle real refraction, edge highlights, and gentle internal reflections.
Inside back wall: a vertical vintage cookbook page (aged off-white, light foxing, realistic paper fibers).
Typography: Elegant serif title set to <InputA> at the top of the page.
The cookbook page may include a small pen-and-ink sketch of the finished dish, but ONLY the sketch is illustrated (everything else in the diorama is photoreal 3D).
Storage (the floating grid):
Small transparent acrylic cubbies adhered to the inner left and right cube walls, floating in mid-air.
Each cubby holds exactly one inferred raw ingredient (selected by the inference engine).
Each cubby has a tiny brown paper tag label with a legible ingredient name (matching the contents).
SCALE (kitchen crew):
A 1:12 scale chef figure stands on the cube floor at the inferred appliance, actively cooking (a believable action pose).
Tiny 1:87 scale construction workers are transporting the hero prop into the cube or lifting miniature crates of the inferred ingredients.
All figures must look like real miniatures (painted resin/plastic texture, tiny seams, realistic scale cues).
Main FOOD (must be realistic 3D, not illustrated):
Place a plated or bowl-served “finished dish” representation of <InputA> inside the cube.
The dish must look edible and authentic: micro-textures, oil sheen, moist surfaces, crisp elements if applicable, natural irregularity, subtle steam wisps.
The finished dish must visually match the inferred ingredient set (no mismatches).
LIGHTING & CAMERA (force photoreal product realism):
Bright, clean studio lighting with softbox diffusion; high clarity; minimal noise; crisp detail.
True-to-life acrylic reflections and refraction without blowing out highlights.
Product photography framing: centered cube, slight 3/4 angle, sharp focus across the diorama (avoid heavy tilt-shift blur).
Photoreal CGI look: physically based rendering (PBR), ray-traced lighting, realistic specular highlights, subsurface scattering where relevant (food).
OUTPUT:
ONE image, 1:1 aspect ratio, high-resolution, high clarity, premium catalog product photography, “miniature life” aesthetic.
STRICT NEGATIVE / EXCLUDE:
illustration-style food, cartoon, anime, painterly, watercolor, flat shading, low-poly, claymation, plastic-looking food, oversmooth surfaces, messy or gibberish text, extra captions outside the cookbook title, cluttered background, heavy depth-of-field blur, motion blur, noise.

3483 / 10000

Based on: Culinary Wonders: Miniature Middle Eastern Feast

Model

Select a model to continue.

Each model exposes different quality and editing capabilities.

Size & Aspect Ratio

Pick the output framing.

Available options change with the selected model.

Select a model to see credit cost. Reference mode adds 1 credit on supported models.

Generate Similar

Generating your image

Image ready

We could not finish this image

Generate Similar

Generating your image

Image ready

We could not finish this image

Send people to the full prompt and image