Text to image3D & miniatures
The exact text-to-image prompt behind this picture, made with Nano Banana 2.1.
Change any box and the prompt below updates. Copy it, or make one with your version.
Top-down photo of a neat hill-shaped arrangement of realistic 3D fridge magnets of Rome, lined up in parallel rows and right angles like a knolling layout. Each magnet is a tiny detailed miniature: a round ancient arena, a domed basilica, a baroque fountain, a pizza slice, a vintage scooter and an espresso cup. At the top centre sits a souvenir magnet printed "ROMA", next to a small handwritten note that reads "Sunny, 24°C". Things that match today's weather are worked into the pile: tiny sunglasses and a gelato cone. No item repeats. Soft even daylight on a white surface.
Prompts stay in English, the language these models follow best. Copy it as it is, or change a few words.
Opens Nano Banana 2.1 with this prompt and the settings below filled in.
91 tested Nano Banana prompts →
The exact price shows before you generate.
lined up in parallel rows and right angles like a knolling layout. Keep that structure and change only the content, or the model loses track of the page.{landmarks}, in the boxes above and keep the rest word for word."ROMA", are printed in the picture. Keep them short and change them to your own.Adapted from a prompt by @NanoBanana, collected in PicoTrex/Awesome-Nano-Banana-images under CC BY 4.0. PopNow rewrote it as a fill-in template and made the example above.
Style prompt guide
A top-down photo of Rome fridge magnets, stacked into a neat hill and laid out in rows like a knolling flat lay, with a tiny gelato cone and sunglasses in the pile. Nano Banana 2.1 works well here because its sharp text rendering keeps the ROMA magnet and the handwritten note readable, while the even daylight keeps every miniature crisp on the white surface.
The style rests on a few visual traits written directly into the prompt. Each one shows in the finished image and changes how the pile reads.
Top-down photo of a neat hill-shaped arrangement of realistic 3D fridge magnets of RomeFixes the camera angle and the core idea: realistic 3D magnets arranged as a hill. Removing "top-down" gives a side view, and the pile loses the flat knolling look that makes the layout easy to read.
lined up in parallel rows and right angles like a knolling layout.Sets the grid that keeps the pile tidy. Without the parallel rows and right angles, the magnets tend to scatter into a loose heap and the knolling style is lost.
a round ancient arena, a domed basilica, a baroque fountain, a pizza slice, a vintage scooter and an espresso cup.Names the Rome items that fill the pile. Cutting the list thins the pile, and changing an item changes the subject, while the tiny detailed miniature wording before it keeps each one small and crisp.
At the top centre sits a souvenir magnet printed "ROMA", next to a small handwritten note that reads "Sunny, 24°C".Adds the only readable words in the scene. The quoted "ROMA" and "Sunny, 24°C" tell the model to write those words exactly; without the quotes the text may come out garbled or missing.
tiny sunglasses and a gelato cone. No item repeats. Soft even daylight on a white surface.The sunglasses and gelato match the note's sunny weather. "No item repeats" prevents duplicate magnets, and changing "Soft even daylight" to a warm light would tint the white base and muddy the colours.
Swaps the city landmarks for a person's tools and trades, so the same style works for a personal keepsake or a gift for someone who works in a craft.
Top-down photo of a neat hill-shaped arrangement of realistic 3D fridge magnets of a chef, a potter and a cyclist, lined up in parallel rows and right angles like a knolling layout. Each magnet is a tiny detailed miniature: a rolling pin, a potter's wheel, a bicycle helmet, a loaf of bread, a paintbrush and a pair of gloves. No item repeats. Soft even daylight on a white surface.
Replaces the Rome landmarks with animals and pet items, which suits a pet owner's gift or a pet shop display.
Top-down photo of a neat hill-shaped arrangement of realistic 3D fridge magnets of pets, lined up in parallel rows and right angles like a knolling layout. Each magnet is a tiny detailed miniature: a tabby cat, a golden retriever, a rabbit, a goldfish bowl, a parrot on a perch and a hamster wheel. No item repeats. Soft even daylight on a white surface.
Keeps the same city-souvenir idea for a different city and street scene, useful when you want a matching set of magnets for another destination.
Top-down photo of a neat hill-shaped arrangement of realistic 3D fridge magnets of Paris, lined up in parallel rows and right angles like a knolling layout. Each magnet is a tiny detailed miniature: a café terrace, a bakery window, a street lamp, a flower kiosk, a bistro chair and a bicycle with a basket. No item repeats. Soft even daylight on a white surface.
Keep the 4:5 ratio, which gives the tall framing that suits a hill-shaped pile with the souvenir magnet at the top. Use the text-to-image mode, since the whole scene comes from the words alone. The 1k resolution is fine for testing the layout, but switch to a higher resolution when the small magnet text and the note need to be sharp for print. For a square product shot, change only the ratio and keep the rest of the prompt as it is.
Yes. Replace "Rome" and the landmark list with the landmarks of another city, and keep "realistic 3D fridge magnets" and "Soft even daylight on a white surface" so the style stays the same. Keep the "ROMA" wording only if you still want that name on the souvenir magnet.
Change the note and the props together. For example, replace "Sunny, 24°C" with a rainy note and swap the sunglasses for a small umbrella, so the magnets still match the forecast. Avoid changing only one of them, because the pile will look mismatched.
The model is built for sharp text, so the ROMA magnet usually comes out legible at this size. Check the small letters on the handwritten note first. If they blur, shorten the note and keep the same style, then test again before using the image.
Your photo
