How to Create a Nano Banana Prompt: Practical Tips and Examples

If you’ve been in the social media space recently, you’ve probably already heard the term “Nano Banana.” Without context, it sounds playful, like something straight out of a cartoon.

However, as most designers, social media managers, and content creators will tell you, Nano Banana is far from that. It’s a powerful AI image generation and editing model that’s revolutionizing how people create content. Of course, as with any other AI tool, its benefits are not automatic — you have to know which prompts to use to get desired images.

That’s exactly what we’re here to help with. This guide dives into the Gemini 2.5 Flash image model and the Gemini 3 Pro Image Preview model, looking at how to create effective Nano Banana prompts.

Nano Banana vs. Nano Banana Pro

First things first, you need to understand the differences between Nano Banana and Nano Banana Pro to prompt them correctly. Nano Banana is the Gemini 2.5 Flash Image model. Nano Banana Pro, on the other hand, is the Gemini 3 Pro Image Preview model. Basically, Nano Banana Pro is an advanced version of Nano Banana.

Nano Banana is primarily designed for speed. It’s an excellent option if you need help with basic image generation and simple editing tasks, and want quick results.

Conversely, Nano Banana Pro is designed for professional, high-fidelity graphic design. It can handle more complex prompts than Nano Banana and efficiently handle tasks that require advanced reasoning.

How to Structure a Nano Banana Prompt

The key to getting great images is to be as specific as possible. The clearer you describe the kind of image you want, the better Nano Banana’s result will be. A good rule of thumb is to define the following elements:

• Subject: Who you want the image to focus on (e.g., a barista).
• Action: What the subject is doing (e.g., pouring coffee into a bright red mug).
• Context: The where and when the action is happening (e.g., in a busy, brightly-lit cafe in the morning with sunlight seeping through its huge windows).
• Style: Your desired aesthetic (e.g., hyper-realistic or cinematic).

You can go further and specify your preferred camera angles and include any specific props you want in the final image. Below is an example of a prompt and what the AI model generates:

“Create an image of a barista pouring coffee into a bright red mug in a busy cafe, with the early morning sunlight seeping through the cafe’s large windows. Create the image in a hyper-realistic style, making sure to include realistic, regular cafe props like menus.”

Specificity is non-negotiable when prompting Nano Banana. It minimizes guessing, making it easier to get what you’re looking for.

Can You Prompt Nano Banana to Edit an Image?

Yes, absolutely! Nano Banana isn’t just an image generator; it is also a powerful reasoning-based AI image editor. You can upload an existing photo and use natural language text prompts to change specific details.

However, editing an image requires a completely different prompting strategy than creating a new one from scratch. If you just type “a person in a red shirt,” the AI might generate a completely new person. To edit like a pro, you need to tell the AI exactly what to keep and what to change.

Here is how you should structure your prompt to edit an image in Nano Banana:

1. Lock in What Must Not Change

The most important rule in Nano Banana editing is to protect first, then direct. You need to explicitly tell the AI to freeze the parts of the image you want to keep.

Start your prompt with strong constraint phrases like:

• “CRITICAL: Keep the person and background EXACTLY the same.”
• “Keep everything unchanged except…”
• “Preserve the original composition and lighting.”YouTube

(Note: If you are using a tool that has a “brush” or “mask” feature, make sure to highlight only the area you want to change.)

2. Use the 5 “Power Action Words”

When you describe the change, use direct action verbs. Extensive testing shows that Nano Banana understands these five specific action words with the highest accuracy:

• Add (e.g., “Add a pair of sunglasses to the dog.”)
• Change (e.g., “Change the cloudy sky to a golden hour sunset.”)
• Make (e.g., “Make the lighting a bit warmer.”)
• Remove (e.g., “Remove the power lines from the background.”)
• Replace (e.g., “Replace the white top with a black t-shirt.”)

3. Keep It to One Change at a Time

Don’t write a mega-prompt asking the AI to change the clothes, swap the background, and add a dog all at once. Nano Banana works best when you give it a single objective.

A perfect editing prompt looks like this:

“Keep everything the same. ONLY CHANGE: replace the woman’s white top with a black t-shirt. Match original lighting.”

By using this simple formula, you can change outfits, fix backgrounds, or remove objects without distorting the original face or scene!

How to Structure a Nano Banana Pro Prompt

As with Nano Banana, it’s important to be as specific as possible when writing your prompt. Nano Banana Pro can handle more complex requests than its preceding model, so don’t be afraid to include many details in your prompt.

Beyond adding elements like the subject, action, and context to prompts, you can also request:

• Text rendering: You can ask the model to add text to your image in a specific font, with specific placement instructions.
• Multi-subject scenes: The model can effectively generate images of multiple subjects, each with distinct actions and contexts. For example, you can ask it to: “Create a skincare product ad’s photoshoot scene. Include a young female model in a mint-green dress, applying the product to her face, with a bottle labeled “ARSENI” beside her. Also include two slightly out-of-focus photographers taking photos of her.”
• Camera specifications: Pro can interpret details like your camera lens specifications. For example, you can ask it to create the effect of using a 10mm–14mm lens if you want your image to have a wide-angle perspective.
• Fine details: The model can effectively handle micro-level details such as fabric and skin textures.
• Multi-image combinations: Nano Banana Pro can combine up to 14 images. If you want to create concept art or a collage, upload the images you want the model to use as reference and provide a descriptive prompt. For example, if you want to create a travel mood board for your friend group, you could upload your images and a photo of your target destination, and ask the model to “Create a photorealistic image of these people smiling at this place.”

Nano Banana Pro: Prime Examples

The Hyper-Realistic Portrait

Focuses on skin texture, emotional depth, and cinematic optics.

Prompt: A cinematic close-up portrait of an elderly craftsman in his cluttered clock workshop. His face is deeply lined with rich textures, captured in sharp detail. Warm, directional light spills from a single desk lamp, creating a dramatic chiaroscuro effect. Dust particles are visible in the light beams. Shot on 35mm film, f/1.8, shallow depth of field with a soft bokeh background of golden gears and springs.

Commercial Product & Typography

Highlights the model’s industry-leading ability to render precise, sharp text on objects.

Prompt: High-end minimalist product photography of a sleek matte black coffee thermos. The thermos is centered on a raw concrete pedestal against a dark charcoal background. The side of the thermos features the word “GeeLark” in a clean, white sans-serif font. Subtle steam rises from the top. Sharp shadows, high contrast, 8k resolution, professional studio lighting.

Surrealist Concept Art

Tests the model’s logical consistency and ability to merge “impossible” elements.

Prompt: A surreal landscape where a giant, transparent lightbulb is partially buried in a desert of white sand. Inside the glass bulb, a lush miniature rainforest grows with a tiny waterfall. A small astronaut stands on the sand outside the bulb, looking in with wonder. The sky is a deep twilight purple with two moons. Ethereal glow, vibrant colors, intricate details.

3D Isometric Character Design

Perfect for generating high-quality assets, icons, or “toy-like” aesthetics.

Prompt: A cute 3D isometric room design representing a “Cat’s Library.” The walls are lined with floor-to-ceiling bookshelves filled with tiny colorful books. A fluffy calico cat is napping on a miniature velvet sofa. Soft, clay-like textures, pastel color palette (mint green and peach). Rendered in Octane, soft global illumination, toy-like aesthetic, high gloss finish.

From NanoBanana Pro Prompt to Automating Posts

Now your NanoBanana Pro images look amazing. But making the picture is only the first step. To get real views, you usually want to turn that image into a short video and post it on platforms like TikTok or Instagram.

Doing this the normal way takes forever. You have to jump between different AI websites, download heavy files, and manually post them from your phone. Plus, if you log into 50 different accounts from the same phone, social media apps will catch you and ban your accounts.

What is GeeLark?

GeeLark is an all-in-one platform that fixes these problems. It gives you “cloud phones.” These are real mobile phones that live on the Internet. To TikTok or Instagram, each cloud phone looks like a totally separate, real person.

GeeLark has powerful AI tools built right into it. You can create your images, turn them into videos, and post them to hundreds of accounts without ever leaving the app.

Step 1: Generate Your Image

Inside GeeLark, you can generate images and videos using its built-in AIGC tools.

Go to Image generator tab, pick NanoBanana Pro, enter your prompt, choose the size, and hit Submit. GeeLark saves the result in your Library so it’s ready for the next step.

Step 2: Turn Your Image Into a Video

Next, you can bring that picture to life. Without leaving the page, simply click over to the Image to video tab.

Select the image you just made, and pick a video model from the drop-down menu, like Veo 3.1 or KLING 2.5. You can choose to generate sound. Hit submit, and in a few minutes, you will have a cool, moving video ready to share.

Step 3: Post Everything on Autopilot

Now that your video is in your Library

GeeLark has an Automation feature. You simply pick a template (like “TikTok video posting“), choose which cloud phones you want to use, and set a time. GeeLark will automatically push your new video to 10, 50, or even 100 accounts. Because it runs in the cloud, it does all the hard work for you — even while you sleep.

Create Descriptive Nano Banana Prompts

Nano Banana can be a valuable asset in your social media journey. With its ability to create and edit images in seconds, it can save you a lot of time and make it easier to stay consistent.

Of course, your prompts will determine just how valuable the tool is. For great images, be as specific as possible and use natural language when writing your prompts.

With GeeLark, you can leverage Nano Banana Pro’s image-generation capabilities without leaving the platform. GeeLark is a cloud phone solution that not only helps manage multiple social media accounts but also automates content posting.

Beyond that, it facilitates seamless image generation by providing quick access to Nano Banana Pro. This means you can create (and edit) your images and schedule them to go live at specific times across multiple accounts, all from the same platform.

FAQs