Nano Banana Prompts: 40 Copy-Paste Edits (2026)

Straight answer: Nano Banana is Google's nickname for its Gemini image model (Gemini 2.5 Flash Image, and now Gemini 3 Pro Image, known as Nano Banana Pro), and the fastest way to get good results out of it is to copy a specific, detailed prompt rather than write a vague one from scratch. Below are 40 prompts organised by job, plus one honest bit nobody selling these lists wants to say out loud: half of them will need a second go before they work.

What Nano Banana is (and why the name matters)

Nano Banana isn't a separate app. It's the codename Google's own team used internally for the Gemini 2.5 Flash Image model, and the name stuck once people started sharing edits on social media. You use it inside the Gemini app, inside Google AI Studio, or inside Whisk. Google then shipped a stronger version, Gemini 3 Pro Image, which everyone is still calling Nano Banana Pro because the old name is more memorable than the new one. Pro handles text on images far better than the original and can blend up to 14 reference images into one edit, which is the bit that changes what small business owners can do with it.

I mention this because half the "Nano Banana prompt" lists floating about were written for the older, weaker model. If your text keeps coming out garbled or your composite images look glued together, check which model you're running before you blame the prompt.

How to use these prompts (five steps, not one)

• Open Gemini, choose the image model (Nano Banana or Nano Banana Pro if it's available on your plan), and upload your source photo.
• Paste the prompt exactly, then swap out anything in square brackets for your own detail.
• Read the result before you accept it. Check hands, check text, check reflections in glass or metal.
• If something's off, don't start again. Reply in the same chat with a correction: "keep everything the same but fix the left hand" works better than a fresh prompt.
• Download at full resolution, not the thumbnail preview, especially if you're putting this on a product listing or a printed flyer.

That correction step is the one people skip. Nano Banana keeps the conversation in memory, so nudging an existing image gets you further than re-rolling a new one from scratch five times in a row.

A worked example: the Etsy candle shop

Say you run an Etsy shop selling soy candles from your kitchen table. You've got one decent phone photo of a candle on a wooden board. You want six listing images without buying a lightbox or hiring anyone.

Prompt one: "Take this candle photo and place it on a white marble surface with soft studio lighting from the top left, keep the candle's label and colour exactly as shown, add a subtle shadow." Prompt two, same image: "Now show this candle lit, with a warm glow from the flame reflecting on the jar, background stays out of focus." Prompt three: "Create a flat lay version of this candle surrounded by dried eucalyptus and a folded linen cloth, autumn colour palette." You now have three usable listing photos from one snap, and you didn't touch a single external editing tool.

The uncomfortable bit: Etsy's own policies require AI-generated or AI-edited product photos to represent the real item. If your candle looks nothing like the marble-surface version once it lands on someone's doorstep, you'll get returns and bad reviews, not sales. Use these prompts to make an accurate photo look better, not to invent a product that doesn't exist.

Product photography prompts

• "Place this product on a plain white background, studio lighting, soft shadow underneath, no other changes to the product itself."
• "Show this product held in a hand, natural skin tone, casual daylight setting, product stays exactly as photographed."
• "Create a lifestyle shot of this product on a kitchen counter with morning light coming through a window, keep the product unchanged."
• "Generate three angles of this product (front, three-quarter, side) using the same lighting and background as the original photo."
• "Add this product into a flat lay with [ITEM 1] and [ITEM 2] on a neutral linen background, top-down view."

Headshot and personal branding prompts

• "Turn this selfie into a professional headshot with a plain grey studio background, soft even lighting, keep my face exactly as it is."
• "Change the background of this photo to a modern office with soft bokeh, keep my clothing and expression unchanged."
• "Give this headshot a warm, natural colour grade like it was shot on a 50mm lens at f/1.8, don't alter my features."
• "Crop and reframe this photo for a LinkedIn profile picture, square format, my face centred and shoulders visible."
• "Replace the background of this photo with a bookshelf in soft focus, keep lighting on my face consistent with the original."

Social media and ad graphic prompts

• "Turn this product photo into a square Instagram post with the words '[YOUR OFFER]' in bold sans-serif text across the top third."
• "Create a carousel-style image showing this product from three different uses, consistent background and lighting across all three."
• "Design a quote card using this background photo with the text '[YOUR QUOTE]' centred in white text with a subtle dark overlay for readability."
• "Turn this photo into a before-and-after style split image, original on the left, [DESCRIBE CHANGE] on the right."
• "Generate a Facebook ad banner, 1200 by 628, using this product photo with a clear call to action button graphic reading 'Shop Now'."

Real estate and interiors prompts

• "Brighten this room photo, correct the white balance, straighten any tilted lines, keep all furniture and layout unchanged."
• "Remove the clutter from this kitchen countertop, keep the cabinets, appliances and layout exactly as shown."
• "Stage this empty room with a simple grey sofa, coffee table and rug, neutral modern style, keep the walls and windows as they are."
• "Change the sky in this exterior house photo to a bright blue sky with a few clouds, keep the building and landscaping unchanged."
• "Show this room with the walls repainted in soft sage green, keep flooring, furniture and lighting the same."

Restoring and enhancing old photos

• "Restore this old, faded photograph, fix the scratches and discolouration, keep the people and background as close to the original as possible."
• "Colourise this black and white photo with natural, realistic tones, don't invent details that aren't visible in the original."
• "Sharpen and upscale this low-resolution photo without changing anyone's face or the background details."
• "Remove the creases and tears from this scanned photograph, keep the composition exactly the same."
• "Fix the red-eye and correct the exposure in this old flash photo, keep everyone's features unchanged."

Removing and replacing backgrounds

• "Remove the background from this photo and replace it with a plain white background, keep the subject's edges clean."
• "Replace the background of this photo with a blurred outdoor park scene, keep the subject's lighting consistent with the new background."
• "Take this photo and place the subject into a studio setting with a solid navy blue backdrop."
• "Remove the person in the background of this photo, fill in the space naturally to match the rest of the image."
• "Replace the sky in this photo with a golden hour sunset, keep the foreground exactly as it is."

Merging and combining images

• "Combine this product photo with this background photo so the product looks naturally placed on the surface, matching the lighting direction."
• "Merge these two portrait photos into a single group photo, matching the lighting and colour tone across both people."
• "Take the outfit from this photo and put it on the person in this second photo, keep the second person's face and pose unchanged."
• "Combine this logo with this product photo, placing the logo subtly on the packaging in a way that looks printed, not pasted."
• "Use these three photos to build one wide banner image, consistent lighting and colour grade across all three."

Style and mood transformations

• "Convert this photo into a watercolour illustration style, keep the composition and subject positions the same."
• "Give this photo a warm, cinematic colour grade like a film still, don't alter the content of the image."
• "Turn this photo into a flat, minimalist vector-style graphic suitable for a website header."
• "Apply a moody, dark academia aesthetic to this photo, deep shadows, warm amber tones, keep the subject recognisable."
• "Recreate this photo in the style of a 1970s film photograph, grain, warm faded colours, keep the composition identical."

The part these listicles don't tell you

Copy and paste prompts get you 70 percent of the way there and then leave you stuck exactly where everyone else who copied the same list gets stuck. If you run a small clothing brand and use prompt 23 word for word, so does the shop three doors down, and your product shots start to look identical to theirs. The prompts here are a starting point, not a finished asset. Change the lighting description, change the surface, change the mood word, and your output stops looching like a template.

There's also a text problem people don't like admitting. Even Nano Banana Pro, which is better at rendering words on images than the original model, still gets small print, fine detail on packaging, and long sentences wrong more often than the marketing demos suggest. If you're putting a Nano Banana image straight onto a product label, a menu, or anything a customer will read closely, check every letter before you print it. I've seen enough AI-generated graphics with a misspelled word sitting proudly in the header to know this isn't rare.

And a legal point worth ten seconds of thought: if you're editing photos of real people who aren't you, or generating images that make a product look different from how it is, that's on you, not on Google. Check platform rules (Etsy, Amazon, Meta ads all have their own AI content policies) before you publish.

When to stop doing this yourself

If you're spending more than an hour a week fighting with prompts to get one usable image, that hour has a cost, and it's usually more than what it would take to build a real repeatable workflow once and reuse it. Businesses I talk to about setting up AI image and content systems, rather than one prompt at a time, tend to get their time back within a fortnight. If that's where you're at, it's worth reading through what working with an AI implementation coach involves before you keep trading hours for images.

I keep every related walkthrough in the ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps. To write your own, try the free tool that writes the prompt for you.

Frequently asked questions

Is Nano Banana the same as Gemini?

Nano Banana is the nickname for Google's Gemini image generation and editing model, not a separate product. You access it through the Gemini app, Google AI Studio, or Whisk, and the newer Nano Banana Pro runs on the Gemini 3 Pro Image model with better text rendering and the ability to blend up to 14 reference images.

Do Nano Banana prompts need to be long?

Specific beats long. A short prompt that names the lighting, background, and exactly what should stay unchanged will outperform a rambling paragraph, because Gemini tends to follow the clearest, most concrete instruction it's given rather than the most detailed one.

Why does my edited photo look different from what I asked for?

Nano Banana sometimes changes small details it wasn't asked to touch, especially faces, hands, and fine text. The fix isn't a new prompt, it's a correction in the same chat, telling it exactly what to keep and what to change, since the model retains context from the conversation.

Can I use Nano Banana images for my Etsy or Amazon listings?

Yes, as long as the image still accurately represents the real product a customer will receive. Both platforms have AI content policies that focus on accuracy, not on banning AI editing outright, so use it to make a true photo look sharper rather than to invent details the product doesn't have.