Gemini AI Photo Prompts: Copy-Paste Guide for Stunning Edits

Gemini can edit photos as well as it writes text, but most people never get past "make this look better" and wonder why the result is disappointing. The trick is not a secret setting. It is the prompt. Gemini's image model, the one people nicknamed Nano Banana, does its best work when you tell it exactly what to change and, just as important, what to leave alone.

This guide is a copy-paste library. Below are 17 prompts grouped by the job you are trying to do, from professional headshots to product shots to restoring a faded family photo. Paste one into the Gemini app, attach your image, and adjust the details in brackets. First, the 30-second version of why these prompts work.

How to use these prompts

You can run every prompt here for free inside the Gemini app on the web or your phone. Open Gemini, click the image icon to upload a photo, paste a prompt, and send. The free tier now runs on Nano Banana 2 (Gemini 3.1 Flash Image), Google's current default image model, and it handles most of these edits well, legible text included. For the hardest cases, like dense poster copy or keeping one face steady across several edits, a paid Google AI Pro, Plus or Ultra subscription lets you open the three-dot menu on a generated image and choose "Redo with Pro" to get Nano Banana Pro's output instead. There is no free way to reach Nano Banana Pro in the Gemini app. If the words "Nano Banana" are new to you, our explainer on what Nano Banana AI actually is covers the full model lineup.

Every prompt below follows the same four-part shape. Learn it once and you can write your own:

That last part is the one people skip, and it is the most valuable. Telling Gemini to keep the face, pose, label, or background unchanged is what stops it from quietly redrawing your whole photo.

1. Portrait and headshot prompts

Turn a phone selfie into something you would put on LinkedIn.

Professional headshot from a selfie:

Turn this selfie into a professional corporate headshot. Keep my face, hairstyle and skin tone exactly the same. Dress me in a dark navy blazer over a white shirt, place me against a soft gray studio background, add even softbox lighting, and frame it from the chest up.

Natural portrait retouch:

Retouch this portrait: even out the skin, remove blemishes and stray hairs, and brighten the eyes very slightly so it still looks like natural window light. Do not change my facial features, my age, or my expression.

2. Background swap prompts

Change where a photo was taken without touching the subject.

Replace the background:

Replace the background of this photo with a clean, out-of-focus coffee shop interior. Keep the person, their pose and the fine edges of their hair intact, and match the lighting on them to the warm tones of the new background.

Clean studio backdrop:

Place the subject on a smooth, continuous pure white studio background with a soft, natural shadow beneath them. Keep the subject, their pose and their colors exactly as they are.

3. Outfit and scene prompts

Restyle clothing or the whole setting while the person stays put.

Change the outfit:

Change my outfit to a fitted charcoal three-piece suit with a burgundy tie. Keep my face, pose, body proportions and the background exactly the same.

Change the season:

Redraw this street scene in autumn: turn the green leaves golden and orange, add a scattering of fallen leaves on the ground, and give the light a warm late-afternoon glow. Keep the buildings and the people in the same positions.

4. Art style transfer prompts

Keep the composition, change the medium.

Watercolor painting:

Reimagine this photo as a soft watercolor painting with visible paper texture and gentle color bleeds, while keeping the composition and the subject clearly recognizable.

Anime illustration:

Convert this portrait into a cel-shaded anime illustration with bold clean outlines and flat color. Keep the person's hairstyle, pose and expression.

5. Photo restoration and enhancement prompts

This is where Gemini genuinely shines, because it repairs based on what the photo already contains.

Restore an old photo:

Restore this old, faded photograph. Repair the scratches and creases, remove the yellow color cast, gently sharpen the faces, and bring back natural contrast. Do not invent any details that are not already in the original.

Colorize black and white:

Colorize this black-and-white photo with realistic, period-accurate colors. Keep the skin tones natural and avoid over-saturation.

Sharpen a blurry shot:

Sharpen and enhance this slightly blurry photo. Recover fine detail in the eyes and hair, reduce noise, and keep it looking like a real photograph rather than an over-processed one.

6. Product and e-commerce prompts

Get a clean catalog shot or a lifestyle scene from one flat photo.

Lifestyle scene:

Place this product into a realistic lifestyle scene on a marble kitchen counter, with a blurred plant in the background and soft morning light. Keep the product's shape, label and colors sharp and unchanged.

Logo mockup:

Put this logo onto a plain white ceramic coffee mug sitting on a wooden table. Wrap the logo around the mug's curve with correct perspective, and add a soft, natural shadow.

7. Composite and text prompts

Blend images or add real, legible words. Nano Banana 2 already renders legible text well on the free tier, so try these prompts there first. If you have a paid Google AI plan, redo the image with Pro when you need the sharpest possible typography.

Combine two photos:

Combine these two images: place the person from the first photo into the landscape from the second. Match the lighting direction and color temperature so the result looks natural, and scale the person realistically for the scene.

Add an object to a person:

Add a pair of round tortoiseshell glasses to this person's face. Match the perspective and the lighting, and add a subtle shadow where the frames meet the skin.

Poster with headline text:

Create a minimalist event poster on a deep navy gradient background. Add the headline "Summer Design Summit" in a bold sans-serif at the top and the date "August 24, 2026" in smaller text at the bottom. Keep the text sharp and correctly spelled.

Social thumbnail:

Turn this photo into a bright, high-contrast thumbnail. Boost the colors, add a subtle glow around the subject, and leave the right third empty for a headline. Keep the face clear and sharp.

Three mistakes that ruin a good edit

• Asking for too much at once. If an edit keeps going wrong, split it. Change the background in one prompt, then the outfit in the next. Gemini handles single, clear changes far better than a five-part instruction.
• Forgetting to protect the subject. Without a "keep my face unchanged" line, the model may redraw features you wanted to preserve. Always name what should stay identical.
• Being vague about style. "Make it nice" gives you the model's guess. "Soft morning light, muted colors, shallow depth of field" gives you a direction.

If you want a broader view of the field before you commit to one tool, our roundup of AI art and image generators compares Gemini's image model against Midjourney, Adobe Firefly and Ideogram side by side.

Frequently asked questions

Are these Gemini photo prompts free to use?

Yes. Every prompt in this guide runs inside the free tier of the Gemini app, on Nano Banana 2, Google's current default image model. Paid Google AI subscribers get a separate daily allotment of the higher-end Nano Banana Pro for tougher edits; once that runs out, their account drops back to the same Nano Banana 2 that free users already have, not to an older or worse model. Google has not published an exact daily number for the free tier.

Why does Gemini keep changing my face?

Because you did not tell it not to. Add an explicit line like "keep my face, hairstyle and skin tone exactly the same" to every portrait edit. For the strongest identity preservation, such as holding one person's face steady across several blended images, Nano Banana Pro does it best, though reaching it now takes a paid Google AI subscription. The free Nano Banana 2 default still respects a clear "keep my face unchanged" instruction on ordinary single-image edits.

What is the difference between Gemini and Nano Banana?

They are the same thing. "Nano Banana" is the nickname that stuck to Google's Gemini image models after one appeared anonymously on a leaderboard. When you edit a photo in Gemini, Nano Banana is the model doing the work. See our full Nano Banana explainer for the model lineup.

Can Gemini add readable text to an image?

Yes. The free-tier default, Nano Banana 2, ships with precision text rendering, which is why the poster and mockup prompts above work without upgrading. For the most demanding typography, such as long paragraphs or multiple languages in one image, Nano Banana Pro still renders text more accurately, but reaching it now requires a paid Google AI Pro, Plus or Ultra subscription.

Why won't Gemini edit a photo of a real person?

Google applies safety rules to edits involving identifiable people, and every image it makes carries an invisible SynthID watermark so it can be flagged as AI-generated later. Ordinary retouching and restyling of your own photos is fine; edits that could be used to impersonate or mislead may be refused.