Skip to main content
Back to Blog
Nano BananaGoogle GeminiAI image generationImagenimage prompts2026

50 Best Nano Banana Prompts in 2026: Copy-Paste Templates That Actually Work

50 copy-paste Nano Banana prompts for Google's image model — character-consistent edits, photo retouching, multi-image composition, scene generation, and text-in-image. Tested templates.

SurePrompts Team
May 6, 2026
31 min read

TL;DR

Fifty Nano Banana copy-paste prompts that exploit what Google's image model is best at — precise instruction-following on edits, character consistency across multiple frames, multi-image composition, accurate text rendering, and photorealistic scene generation. Organized by use case so designers, marketers, and creators can paste and ship.

Nano Banana — Google's image generation and editing model in the Gemini 2.5 family — isn't trying to beat Midjourney at moody atmospheric portraits. It's winning a different game entirely: precise instruction-following on edits, character consistency across frames, multi-image composition, and text rendering that actually spells things right. The 50 copy-paste Nano Banana prompts below exploit these strengths directly. The problem is that most people prompt it like Midjourney, which means they get mediocre results and write it off — when really they just need to speak its language.

What Nano Banana Is Best At

Precise instruction-following on edits. Tell Nano Banana to keep the subject identical and change only the background, and it will do that — without accidentally reshaping your subject, shifting the lighting on their face, or inventing new elements you didn't ask for. This is the model's clearest strength compared to diffusion-only alternatives. The key is being explicit about both what changes and what stays the same. Most people only say what changes. That's half the instruction.

Character consistency across multiple frames. Attach a reference image, describe what should be different in the new frame (new outfit, new setting, new pose), and the model preserves the recognizable core of the character — facial structure, hair, proportions. This matters enormously for creators building visual stories, marketers running campaigns with a recurring character, or anyone producing a sequence of images that needs to feel cohesive. For a deeper breakdown of how to structure reference-based prompts, see the AI image prompting complete guide.

Multi-image composition. Nano Banana can take two or three source images and combine them into a single scene — a subject from one photo placed believably into the environment from another, with matched lighting direction and perspective. This is genuinely useful for product photography, editorial work, and visual storytelling, and it removes the need for manual compositing in many cases.

Text rendering. This is where most image models trip up. Nano Banana, drawing on the broader Gemini family's language foundations, handles spelled-out text in images significantly better than earlier image generation models — especially for short, specific text like headlines, signs, labels, and poster copy. Putting the exact text in quotation marks in your prompt is the critical step that most people skip.

Photorealistic scene generation with attention to physics and lighting. When you describe a scene in natural language with specific lighting details — golden hour coming from the left, soft window light diffused through linen — the model takes those cues seriously. It doesn't flatten everything into generic "photorealistic" sameness. Specificity in lighting and material descriptions produces noticeably better results.

50
Copy-paste Nano Banana prompts across editing, character consistency, composition, scenes, and text-in-image

Precise Photo Editing Prompts (1–8)

1. Background Swap

code
[Reference image: attach the original photo]

Edit instructions: Keep the subject exactly as shown — same pose, same
expression, same clothing, same hair, same lighting on their face and body.
Replace only the background with [DESCRIBE NEW BACKGROUND — e.g., a minimalist
white studio backdrop / a sunlit Kyoto street in autumn / a modern glass-walled
office]. The new background should match the lighting direction already on the
subject. Do not alter any part of the subject.

2. Object Removal

code
[Reference image: attach the source photo]

Edit instructions: Remove [SPECIFIC OBJECT — e.g., the lamp in the left corner /
the coffee cup on the table / the person in the background on the right] from
this image. Fill the area naturally using the surrounding context — match the
floor, wall, or surface texture that would logically be there. Keep everything
else in the image identical: no cropping, no changes to other subjects, no
color shift on the rest of the scene.

3. Object Addition

code
[Reference image: attach the source photo]

Edit instructions: Add [NEW OBJECT — e.g., a small potted succulent / a matte
black laptop / a steaming mug of coffee] to [SPECIFIC LOCATION — e.g., the
right side of the desk / the foreground on the lower left]. The added object
should match the lighting direction and color temperature of the existing scene.
Scale it realistically. Do not change anything else in the image.

4. Color Change on Specific Element

code
[Reference image: attach the source photo]

Edit instructions: Change the color of [SPECIFIC ELEMENT — e.g., the car /
the sofa / the jacket the person is wearing] from [CURRENT COLOR] to
[NEW COLOR — e.g., deep forest green / matte terracotta / soft ivory]. Keep
every other detail exactly the same: the shape, the texture, the shadows, the
reflections, and all other objects in the frame. Do not alter any other colors
in the image.

5. Lighting Adjustment

code
[Reference image: attach the source photo]

Edit instructions: Adjust the lighting in this image to look like [TARGET
LIGHTING — e.g., golden hour with warm directional light coming from the left /
cool overcast daylight / dramatic single-source side lighting]. Keep the subject,
composition, and all objects identical. Change only how the light falls on the
scene — adjust shadows, highlights, and color temperature accordingly. The
overall framing should remain the same.

6. Expression Edit

code
[Reference image: attach the portrait photo]

Edit instructions: Change the subject's expression from [CURRENT EXPRESSION —
e.g., neutral / serious] to [NEW EXPRESSION — e.g., a natural, relaxed smile /
confident and direct]. Keep their face structure, skin tone, hair, clothing, and
the background completely unchanged. Only the expression and the subtle muscle
movements that accompany it should change. The result should look like a natural
photo of the same person, not a composite.

7. Season Change

code
[Reference image: attach the landscape or outdoor scene photo]

Edit instructions: Transform this scene from [CURRENT SEASON — e.g., summer] to
[TARGET SEASON — e.g., late autumn / deep winter with snow / early spring with
blossoms]. Adapt the foliage, ground cover, sky color, and ambient light to
match the new season. Keep the architecture, roads, composition, and any
permanent structures identical. The scene should look like a photograph taken
from the same spot in a different time of year.

8. Time-of-Day Shift

code
[Reference image: attach the exterior or landscape photo]

Edit instructions: Shift this scene from [CURRENT TIME — e.g., midday] to
[TARGET TIME — e.g., blue hour just after sunset / pre-dawn with deep navy sky /
golden hour late afternoon]. Adjust the sky, ambient light, shadows, and
artificial light sources (if any) to match the new time of day. Keep all
structures, objects, and composition exactly as shown. The image should feel
like the same location photographed at a different hour.

Character & Subject Consistency Prompts (9–15)

9. Same Character, New Outfit

code
[Reference: attach a clear photo of the character or person]

Generate a new image of this character. Keep their face, skin tone, hair color
and style, and overall build consistent with the reference — they should be
immediately recognizable as the same person. New outfit: [DESCRIBE OUTFIT IN
DETAIL — e.g., a tailored charcoal suit with a white shirt, no tie]. Setting:
[DESCRIBE — e.g., standing in a modern lobby, neutral background, soft
overhead lighting]. The character should appear to be the same person wearing
different clothes, not a different person.

10. Same Character, New Pose

code
[Reference: attach a photo of the character]

Generate a new image of this character in a different pose. Keep their
appearance — face, hair, proportions, skin tone — consistent with the reference.
New pose: [DESCRIBE POSE — e.g., seated at a desk, leaning slightly forward,
looking at camera with a direct expression / standing with arms crossed,
three-quarter angle, relaxed]. Clothing: keep the same as the reference unless
specified. Lighting: [DESCRIBE — e.g., soft natural window light from the left].

11. Same Character, New Scene

code
[Reference: attach a photo of the character]

Generate a new image of this character placed in a new environment. Keep their
appearance consistent: same face structure, hair, skin tone. New scene:
[DESCRIBE ENVIRONMENT IN DETAIL — e.g., walking through a rain-wet Tokyo
alley at night, neon reflections on the pavement, medium shot, shallow depth of
field]. The character's scale and perspective should fit naturally into the
environment. They should look like they belong in the scene, not pasted into it.

12. Same Product, New Angle

code
[Reference: attach a product photo]

Generate a new image of this exact product from a different angle. Keep the
product's color, texture, shape, branding, and details identical to the
reference. New angle: [DESCRIBE — e.g., top-down flat lay / 45-degree front
isometric / close-up of the top third]. Surface: [DESCRIBE — e.g., white marble /
raw oak wood / matte concrete]. Lighting: [DESCRIBE — e.g., soft diffused
studio light, no harsh shadows]. The product should look identical to the
reference — just repositioned.

13. Same Person, Aged

code
[Reference: attach a portrait photo]

Generate a new portrait of this person aged approximately [NUMBER] years. Keep
their core facial structure, eye color, and distinguishing features recognizable.
Add realistic age indicators: [SPECIFY — e.g., silver-gray hair, deeper lines
around the eyes and mouth, slight softening of the jawline]. Keep their
expression [DESCRIBE — e.g., calm and confident]. The background should be
[DESCRIBE — e.g., simple and neutral]. The result should look like a natural
photograph, not a digital effect.

14. Multi-Frame Story (3 Frames, Same Character)

code
[Reference: attach a photo of the character]

Generate three separate images that form a visual sequence featuring the same
character. Keep their appearance fully consistent across all three — same face,
hair, and build.

Frame 1: [DESCRIBE SCENE AND ACTION — e.g., the character sitting alone at a
café table, looking thoughtful, morning light]
Frame 2: [DESCRIBE SCENE AND ACTION — e.g., the same character mid-conversation,
leaning forward, engaged expression, same café]
Frame 3: [DESCRIBE SCENE AND ACTION — e.g., the character leaving the café,
smiling, afternoon light through the door]

Each frame should feel like it was taken at a slightly different moment in
the same story.

15. Brand Mascot Consistency

code
[Reference: attach the mascot illustration or image]

Generate a new image of this brand mascot in a new situation. Keep the mascot's
visual identity completely consistent with the reference: same color palette,
same proportions, same face design, same distinctive details [LIST KEY FEATURES —
e.g., the orange beak, the round white body, the small black eyes]. New scenario:
[DESCRIBE — e.g., the mascot is holding a smartphone and waving at the camera /
the mascot is running, viewed from the side in a dynamic pose]. Style: match
the illustration style of the reference exactly.

Multi-Image Composition Prompts (16–22)

16. Combine Subject and Environment

code
[Reference image 1: attach photo of the subject — person or object]
[Reference image 2: attach photo of the environment or location]

Create a new image that places the subject from image 1 into the environment
of image 2. Match the lighting direction and color temperature from the
environment image. The subject's scale should look natural relative to the
environment. Adjust shadows and reflections so the subject appears to exist
in the scene rather than being composited onto it. Keep the subject's
appearance identical to image 1.

17. Swap Subjects Between Two Photos

code
[Reference image 1: attach the source photo with the subject you want to move]
[Reference image 2: attach the destination scene]

Take the [SUBJECT — e.g., the person / the dog / the vintage car] from image 1
and place them into the scene of image 2. Remove any gap or inconsistency at the
extraction point in image 1 naturally. In image 2, position the subject at
[LOCATION IN FRAME — e.g., the center foreground / left middle ground] with
scale and perspective that fits the scene. Match lighting from image 2.

18. Blend Two Products Into One Scene

code
[Reference image 1: attach product A]
[Reference image 2: attach product B]

Create a single product photography image that features both products together.
Place [PRODUCT A] on the left and [PRODUCT B] on the right, or arrange them
[DESCRIBE DESIRED ARRANGEMENT — e.g., one in foreground one in background /
both at equal depth side by side]. Surface: [DESCRIBE — e.g., textured linen /
dark slate / glass shelf]. Lighting: [DESCRIBE — e.g., soft diffused white
studio light]. Both products should look exactly as they appear in their
reference images.

19. Add Person to Existing Photo

code
[Reference image 1: attach the existing scene photo]
[Reference image 2: attach a photo of the person to add]

Add the person from image 2 into the scene of image 1. Place them at
[LOCATION — e.g., standing in the background on the right / seated at the
empty chair in the foreground]. Match their scale and perspective to the scene.
Their lighting should match the ambient light direction and color temperature
of image 1. Their presence should look natural — as if they were there when
the original photo was taken. Keep the rest of image 1 unchanged.

20. Layer Foreground onto Background

code
[Reference image 1: the foreground element — attach photo]
[Reference image 2: the background scene — attach photo]

Composite image 1 as the foreground element over the background of image 2.
The foreground should be positioned at [DESCRIBE PLACEMENT — e.g., bottom
center, filling roughly 40% of the frame height]. Remove any background from
image 1 that would conflict with image 2. Adjust the foreground's lighting
to match the background. Add a subtle shadow or reflection where the foreground
meets the ground plane of image 2 to ground it in the scene.

21. Replace One Object With Another From a Second Image

code
[Reference image 1: attach the base scene photo]
[Reference image 2: attach the photo containing the replacement object]

In image 1, replace [OBJECT TO REPLACE — e.g., the generic white mug on the
table / the plain chair in the corner] with [REPLACEMENT OBJECT — e.g., the
ceramic mug from image 2 / the wooden chair from image 2]. Keep everything else
in image 1 exactly the same. Scale the replacement object to match the
perspective and proportions of the scene. Adjust its lighting to match
the ambient light in image 1.

22. Composite Group Photo

code
[Reference image 1: attach group photo A — missing some members]
[Reference image 2: attach photo of the person or people to add]

Add the person (or people) from image 2 into the group photo in image 1.
Position them [DESCRIBE — e.g., in the back row on the far right / standing
next to the person in the blue shirt]. Match their scale to the other people
in the frame. Lighting and skin tones should blend naturally with the rest of
the group. The final image should look like a single photo taken of everyone
at the same time. Keep all existing people in image 1 unchanged.

Photorealistic Scene Prompts (23–29)

23. Lifestyle Scene

code
A photorealistic lifestyle photo of [SUBJECT — e.g., a woman in her early 30s]
[ACTIVITY — e.g., reading a book on a balcony with a coffee cup nearby]. Setting:
[ENVIRONMENT — e.g., a sun-drenched apartment balcony, lush plants around the
railing, city skyline soft in the background]. Lighting: warm morning light
coming from the right, gentle and natural. Color palette: warm tones, creams,
soft greens. Camera angle: [DESCRIBE — e.g., medium shot from slightly below
eye level]. The scene should feel unhurried and lived-in, not staged.

24. Interior Architecture

code
A photorealistic interior photograph of [SPACE TYPE — e.g., a compact Tokyo
apartment living room / a Scandinavian kitchen / a mid-century modern home
office]. Key design elements: [LIST 3-5 ELEMENTS — e.g., a low platform sofa
in oatmeal linen, a walnut coffee table, a single arc floor lamp, built-in
shelves with minimal objects]. Lighting: [DESCRIBE — e.g., diffused natural
light through sheer curtains on the left wall, no harsh shadows]. Camera angle:
wide shot from one corner of the room, showing depth. No people. The image
should look like it belongs in an architecture magazine.

25. Food Photography

code
A photorealistic food photography shot of [DISH — e.g., a bowl of ramen with
chashu pork, soft-boiled egg, nori, and bamboo shoots / a rustic sourdough loaf,
freshly cut, on a linen cloth]. Surface: [DESCRIBE — e.g., aged oak cutting
board / dark slate with crumbs / white ceramic plate on a linen napkin].
Lighting: [DESCRIBE — e.g., soft natural side light from a window on the left,
warm and slightly low]. Camera angle: [DESCRIBE — e.g., overhead flat lay /
45-degree three-quarter angle]. The steam, texture, and colors should feel
appetizing and true to the real dish.

26. Nature Landscape

code
A photorealistic landscape photograph of [LOCATION TYPE — e.g., a misty
mountain valley at dawn / a coastal bluff at golden hour / a dense old-growth
forest after rain]. Key elements: [DESCRIBE WHAT SHOULD BE IN THE FRAME —
e.g., a winding gravel path leading into the forest, shafts of light through
the canopy, soft fog hanging low]. Lighting: [DESCRIBE — e.g., diffused early
morning light, cool blue-green shadows]. Color palette: [DESCRIBE — e.g.,
muted greens, silver grays, gold accents where light breaks through].
No people, no structures. Camera angle: [DESCRIBE — e.g., wide angle from
ground level, foreground in sharp focus, background softly atmospheric].

27. Urban Environment

code
A photorealistic street photography style image of [URBAN SETTING — e.g.,
a narrow alley in Osaka at night / a Paris side street on an overcast morning /
a New York corner store at dusk]. Key details: [DESCRIBE — e.g., wet pavement
reflecting neon signs, a few pedestrians blurred in motion, a convenience store
glowing warm in the background]. Lighting: [DESCRIBE — e.g., mixed neon and
fluorescent, cool shadows, warm pools of light on the wet ground]. Camera:
[DESCRIBE — e.g., medium shot, 35mm perspective, slightly grainy feel without
being a filter]. The image should feel like a candid moment, not a render.

28. Automotive

code
A photorealistic automotive photograph of [CAR — e.g., a matte black 1969
Mustang fastback / a modern silver electric sedan]. Setting: [DESCRIBE — e.g.,
parked on a deserted coastal road at dusk / in a minimalist concrete parking
structure with dramatic overhead lighting]. Lighting: [DESCRIBE — e.g., golden
hour from behind and to the right, catching the roofline and rear quarter panel /
cool overhead fluorescent, sharp shadows below the chassis]. Camera angle:
[DESCRIBE — e.g., low three-quarter front view, shooting up slightly]. The
paint, reflections, and chrome should look physically accurate.

29. Fashion Editorial

code
A photorealistic fashion editorial photograph of [SUBJECT — e.g., a tall
person with short natural hair]. Wearing: [DESCRIBE OUTFIT IN DETAIL — e.g.,
an oversized camel trench coat, wide-leg charcoal trousers, white low-top
sneakers]. Setting: [DESCRIBE — e.g., standing in the middle of a wide empty
concrete street, industrial buildings softly out of focus behind]. Lighting:
[DESCRIBE — e.g., overcast daylight, even and shadowless, colors saturated but
not over-processed]. Camera: [DESCRIBE — e.g., medium full-body shot, slightly
below eye level, clean and editorial]. The image should feel like it belongs
in a fashion magazine spread.

Text-in-Image Prompts (30–36)

30. Branded Poster With Headline

code
Generate a vertical poster (portrait orientation) featuring the headline text
"MAKE YOUR MARK" displayed in large, bold type as the primary visual element.
Style: minimal and modern, clean geometric layout. Background: deep navy blue.
Text color: bright white for the headline. Below the headline, smaller text
reading "Spring Collection 2026" in a light-weight sans-serif. At the bottom,
the word "BRAND NAME" as a centered footer. Spelling and letter spacing must
be accurate. All text should be legible and precisely placed on the layout grid.

31. Sign With Specific Text

code
Generate a photorealistic image of a [SIGN TYPE — e.g., hand-painted wooden
café sign / brushed metal storefront sign / vintage enamel road sign] displaying
the exact text "[YOUR SIGN TEXT — e.g., OPEN FROM 8 / LOST CREEK TRAIL — 2.4 MI /
COFFEE & GOOD COMPANY]". Style: [DESCRIBE — e.g., weathered and charming, painted
in cream on dark green / clean and modern, laser-etched on matte steel]. The
sign should be photographically realistic, mounted on [DESCRIBE CONTEXT — e.g.,
a whitewashed brick wall / a wooden post in a forest]. Every letter must be
correctly spelled and evenly spaced.

32. Book Cover With Title

code
Generate a photorealistic book cover for a [GENRE — e.g., literary fiction /
business nonfiction / science fiction] book. Title displayed prominently:
"THE LONG RETURN". Author name below: "A. Reyes". Style: [DESCRIBE — e.g.,
a moody, desaturated photograph of an empty road at night as the cover image,
title in large serif type in white, author name in smaller type below]. Spine
and back cover visible if shown at an angle. The title and author name must
be correctly spelled. The overall design should look like a real published book
from a major press, not a self-published template.

33. Social Media Card With Copy

code
Generate a social media card sized for a square post (1:1 ratio) featuring
the following text, displayed clearly and legibly:

Headline: "5 Things Nobody Tells You About Freelancing"
Subtext: "Thread starts here →"
Account handle: "@yourhandle"

Style: bold typographic layout, high contrast. Background: solid warm coral.
Text: white. Use a clean, modern sans-serif for all text. The hierarchy should
clearly communicate headline first, subtext second, handle third. All spelling
must be exact. No additional decorative elements beyond the typography and a
minimal dividing line between the headline and subtext.

34. Label and Packaging Text

code
Generate a photorealistic product label for a [PRODUCT TYPE — e.g., small-batch
hot sauce bottle / artisan coffee bag / natural skincare serum]. Label must
display the following text accurately:

Product name: "WILDFIRE"
Variant: "Smoked Chipotle"
Net weight: "148 mL / 5 fl oz"
Tagline: "Made in small batches. Heat on purpose."

Label style: [DESCRIBE — e.g., kraft paper texture with deep red ink / minimal
white label with black typography and a single color accent]. The label should
look like a real CPG product ready for retail. All text must be correctly spelled
and legible at product scale.

35. Billboard Mockup

code
Generate a photorealistic billboard mockup showing an outdoor billboard in
[SETTING — e.g., an urban street intersection / along a highway at dusk] displaying
the following advertisement:

Large text: "YOUR PROBLEM. SOLVED."
Smaller text below: "SolveCo — Try Free for 30 Days"
Website URL: "solvecompany.com"

Billboard dimensions: standard landscape billboard proportion. Brand color
background: [DESCRIBE — e.g., electric blue]. Text: white. The billboard structure
should look physically real — metal frame, mounting hardware, slight surface
texture. The surrounding environment should be plausible for the setting. All
text must be correctly spelled.

36. Certificate or Diploma Layout

code
Generate a photorealistic image of a formal certificate or diploma with the
following text displayed accurately:

"Certificate of Completion"
"This certifies that"
"[RECIPIENT NAME]"
"has successfully completed"
"Advanced Brand Strategy — 2026"
"Issued by the Institute for Brand Excellence"
Date: "May 6, 2026"

Style: [DESCRIBE — e.g., formal and traditional — ivory paper, navy and gold
border, serif typography throughout / modern and minimal — white with clean
lines, a single accent color]. The layout should look like a real institutional
certificate. All text must be correctly spelled and formatted with appropriate
typographic hierarchy.

Product & Commercial Prompts (37–43)

37. Product Hero Shot

code
A photorealistic product hero shot of [PRODUCT — e.g., a slim matte black
bluetooth speaker / a glass skincare serum bottle with a white dropper lid].
Background: [DESCRIBE — e.g., pure white / soft gradient from light gray to white /
rich dark charcoal]. Lighting: [DESCRIBE — e.g., soft box light from the upper
left, secondary fill light from the right, subtle shadow below and behind the
product]. Camera angle: [DESCRIBE — e.g., eye level, centered, three-quarter
view / straight-on front]. The product surface should show accurate reflections
and material texture. No text, no props. The product alone, shot for an
e-commerce hero image.

38. Product With Hand Model

code
A photorealistic product photography image showing [PRODUCT — e.g., a ceramic
coffee mug / a lipstick tube / a leather wallet] being held or used by a hand.
Hand: [DESCRIBE — e.g., a well-groomed hand with light skin and short nails /
a hand with medium-dark skin tone, no jewelry]. Grip or interaction: [DESCRIBE —
e.g., held in one hand, product facing camera, 45-degree angle / hand gently
pressing the pump of the bottle, product in focus]. Background: [DESCRIBE — e.g.,
soft white / blurred neutral environment]. Lighting: [DESCRIBE — e.g., even
soft diffused light, no harsh shadows on the hand or product]. The product
should be in sharp focus.

39. Product on Context Surface

code
A photorealistic product photography image of [PRODUCT — e.g., a candle in an
amber glass jar / a premium notebook] placed on [SURFACE — e.g., a marble
bathroom counter with a few dried eucalyptus stems nearby / a raw linen surface
with a pencil and a folded letter card beside it]. Lighting: [DESCRIBE — e.g.,
warm natural window light from the left, creating a soft shadow to the right].
Camera: [DESCRIBE — e.g., low angle, 30 degrees above the surface, slightly
toward the product]. The styling props should complement the product without
competing with it for attention. The product should remain the primary subject.

40. Product Variants — Color Swap

code
[Reference image: attach the product photo in its original color]

Generate three additional product images showing the same product in different
color variants. Keep every other detail exactly the same as the reference:
shape, texture, material finish, lighting, background, and camera angle.
Color variant 1: [COLOR — e.g., sage green]
Color variant 2: [COLOR — e.g., warm terracotta]
Color variant 3: [COLOR — e.g., deep midnight navy]

Each variant should look like an official product photo, consistent with
the reference but in the new color. Output all three as separate images.

41. Product Packaging Visualization

code
A photorealistic image of a [PACKAGING FORMAT — e.g., folded cardboard box /
flexible pouch / glass bottle with custom label] for [PRODUCT TYPE — e.g., a
premium tea brand / a supplement brand / a specialty condiment]. The packaging
displays [DESCRIBE DESIGN — e.g., a minimalist botanical illustration in olive
green and gold on a cream background, with the product name prominently on the
front panel]. Shown from [ANGLE — e.g., a three-quarter isometric view / front
and slightly angled to show depth]. Lighting: soft studio light, subtle
shadows to give it dimension. The result should look like a packaging render
ready for a brand presentation.

42. Product-in-Use Lifestyle

code
A photorealistic lifestyle photo showing [PRODUCT — e.g., a portable Bluetooth
speaker / a stainless steel water bottle / a skincare product] being actively
used by a person in a real environment. Person: [DESCRIBE — e.g., a person in
their late 20s, casual clothing, outdoors]. Setting: [DESCRIBE — e.g., sitting
on a rocky outcrop with mountain views in the background / at a café table,
mid-afternoon light]. Interaction: [DESCRIBE HOW THEY'RE USING IT — e.g.,
holding it while looking at it and smiling / pouring from it into a cup].
Lighting: [DESCRIBE]. The product should be clearly visible and recognizable,
but the overall shot should feel natural and not overly staged.

43. Product 360-Style Angles

code
[Reference image: attach the product photo]

Generate four images of this product from different angles to simulate
a 360-degree view set. Keep the product's appearance identical to the
reference in every image — same color, same texture, same details. Backgrounds
should be consistent across all four: [DESCRIBE — e.g., pure white with a
soft shadow directly below].

Image 1: Front-facing, straight-on, eye level
Image 2: 45-degree angle from the right, slightly elevated
Image 3: Side profile from the right
Image 4: Three-quarter rear view from the left

Each image should look like it was taken in the same studio setup,
just from a different position around the product.

Creative & Conceptual Prompts (44–50)

44. Surreal Composite

code
A photorealistic surreal composite image of [SUBJECT — e.g., a woman in a white
dress] standing in [IMPOSSIBLE ENVIRONMENT — e.g., a library that exists inside
a glass sphere floating in open sky / a forest of giant hourglasses with real
sand falling through them]. The physical details should be rendered with
photographic accuracy — materials, shadows, light behavior — even though the
overall scene is impossible. Mood: [DESCRIBE — e.g., dreamlike but not
unsettling / contemplative and vast]. Color palette: [DESCRIBE — e.g., warm
gold and amber / cool silver and blue]. Camera: [DESCRIBE — e.g., wide shot
showing the full environment with the subject as a small figure within it].

45. Photo-Illustration Hybrid

code
Generate an image that blends photographic realism with illustration elements.
Photographic base: [DESCRIBE — e.g., a realistic portrait of a person / a
real-looking urban scene]. Illustrated overlay: [DESCRIBE ILLUSTRATION ELEMENTS —
e.g., botanical line drawings growing over the background / hand-drawn constellations
connecting points of light in the scene / ink-wash pattern spreading from the
subject's hands]. The illustration elements should feel intentional and designed,
not random. The photo elements should feel photographically real. Style: [DESCRIBE
FEEL — e.g., editorial and sophisticated / playful and printed / fine-art and
considered]. The two elements should coexist in a single cohesive image.

46. Conceptual Editorial

code
A photorealistic conceptual editorial image representing the idea of
[ABSTRACT CONCEPT — e.g., information overload / creative burnout / the
weight of unread messages / quiet ambition]. Visual approach: [DESCRIBE HOW
YOU WANT TO VISUALIZE IT — e.g., a person sitting at a desk buried up to their
shoulders in paper, lit by a single lamp, looking calm / a figure walking
toward a door made entirely of clocks]. Render it photographically — the
concept is communicated through the image itself, not through text overlays.
Mood: [DESCRIBE — e.g., slightly unsettling but composed / poignant and quiet].
Color palette: [DESCRIBE — e.g., desaturated with a single warm light source].

47. Dreamscape

code
A photorealistic dreamscape image of [DESCRIBE CORE SCENE — e.g., a vast
desert at night where the sand is reflective like still water, mirroring an
impossibly detailed starfield above / a dense underwater forest of ancient
trees with light filtering down from a surface far above]. The environment
should be internally consistent — the physics of this world should feel real
even though the world itself is impossible. Lighting: [DESCRIBE — e.g., cool
silver moonlight and bioluminescent accents in blues and greens]. No people
unless specifically desired. Scale should feel immense. The image should feel
like a location that exists somewhere, not a special effect.

48. Story-Illustration Sequence

code
[Reference: attach a reference image of the character or world, if you have one]

Generate three sequential images that tell a short visual story. The same
character and setting should appear consistently across all three.

Image 1 (Setup): [DESCRIBE — e.g., the character notices something unusual —
a light under a door at the end of a long empty corridor. Medium shot, dim
and slightly eerie.]
Image 2 (Tension): [DESCRIBE — e.g., the character reaches out to push the door
open. Close-up on their hand, light spilling through the gap. Anticipation.]
Image 3 (Resolution): [DESCRIBE — e.g., the door opens to reveal a sunlit garden,
unexpected and warm. Wide shot. The character small in the doorway, the garden
vast ahead.]

Keep character, lighting logic, and visual style consistent across all three.

49. Before-After Transformation

code
Generate two images as a matched pair showing a before-and-after transformation.
Both images should be photographed from the identical camera angle and framing.

Before image: [DESCRIBE STARTING STATE — e.g., a neglected backyard with
overgrown grass, a broken fence, scattered debris, and a weathered brick patio /
a small living room with mismatched furniture, cluttered shelves, dim lighting]

After image: [DESCRIBE TRANSFORMED STATE — e.g., the same backyard, now a
landscaped outdoor living space — clean patio, raised garden beds, warm string
lights, a simple wooden dining table / the same room, transformed — minimal
furniture, warm layered lighting, curated shelves, a cohesive color scheme]

The camera angle, time of day, and perspective should match exactly between
both images so they read as a clear before-and-after pair.

50. Cinematic-Style Still

code
A photorealistic cinematic-style still image, as if from a [GENRE — e.g., a
slow-burn thriller / a quiet European drama / a neon-lit action film]. Scene:
[DESCRIBE — e.g., a lone figure standing on a rain-soaked train platform at
2am, a single overhead light above them, the empty tracks stretching to darkness
on both sides / two people sitting across from each other at an outdoor café,
mid-conversation, one looking away]. Lighting: [DESCRIBE — e.g., single
practical overhead light against darkness / warm café glow against cool blue
street]. Aspect ratio feel: wide, cinematic. The image should look like a frame
pulled from a real film — composed with intention, lit with purpose, shot by
a cinematographer who knows what the scene is about.

Nano Banana Power Tips

1

List what STAYS and what CHANGES. For edit prompts, most people only describe what they want changed. Nano Banana responds better when you explicitly list what should remain untouched — "keep the pose, the expression, the clothing, the background" — before describing the edit. This anchors the model and reduces unwanted changes.

2

Attach the reference and describe what makes the subject identifiable. For character consistency prompts, don't just attach the image and hope. Describe the key identifiers: "same face structure, same hair color and texture, same proportions." The model uses both the visual reference and your verbal description — give it both.

3

Tell it how lighting and perspective should match. In composition prompts, be explicit: "match the lighting direction from image 2 — the light is coming from the upper left." If you say nothing about lighting, the model will make a decision for you. It's often a fine decision, but if you need professional-quality compositing, describing the match conditions gets you much closer on the first pass.

4

Put the exact text in quotation marks. For any text-in-image prompt, write the precise text in quotes in your prompt. Don't paraphrase it. Don't describe it. Write it exactly as it should appear. This is the single most reliable lever for getting correctly spelled, accurately rendered text.

5

Skip Midjourney parameter syntax entirely. Nano Banana wants natural language — conversational, instructional sentences. There are no --ar, --style, --chaos parameters here. Writing detailed paragraph descriptions of what you want, as if explaining it to a photographer, produces better results than terse comma-separated style keywords. Write like you're briefing a person, not invoking a command.

6

Iterate by continuing the conversation. Nano Banana through Gemini tracks the conversation context. If the first result is close but not quite right, respond with "Now do the same, but change [specific thing]" rather than starting over from a fresh prompt. This is faster and often more precise than rewriting the whole prompt. Chain small refinements rather than issuing monolithic re-prompts.

Before

Edit this photo to make the background better.

After

Keep the subject exactly as shown — same pose, same clothing, same expression, same lighting on their face. Replace only the background with a sun-drenched outdoor café terrace, golden afternoon light coming from the right side to match the existing light direction on the subject. Preserve every detail of the subject. The new background should look photographically real, not illustrated.

Start Building Better Prompts

These 50 templates work because they tell Nano Banana what to preserve as much as what to change. The model is capable of precise, surgical edits — but only if your prompt is precise. Generic descriptions produce generic results; specific, structured instructions produce output you can actually use.

The AI prompt generator builds structured image prompts like these automatically — describe what you need and get a ready-to-use prompt tuned for instruction-following models. For a complete breakdown of how image prompting works across all major models, read the AI image prompting complete guide. If you're comparing Nano Banana against Midjourney's strengths, the best Midjourney v7 prompts guide is the right follow-up. And for ChatGPT's built-in image generation capabilities, see the ChatGPT image prompts guide.

Try it yourself

Build expert-level prompts from plain English with SurePrompts — 350+ templates with real-time preview.

Open Prompt Builder

Get ready-made Gemini prompts

Browse our curated Gemini prompt library — tested templates you can use right away, no prompt engineering required.

Browse Gemini Prompts