Skip to main content

Mastering Nano Banana: A Complete Guide to Better AI Image Prompts

· 7 min read
Kenny Mário
Kenny Mário
Co-Founder @taktlos, Developer of Cubby Image

Getting great results from AI image generation isn't just about having the right tools—it's about knowing how to communicate with them. After helping countless users work with Nano Banana (Google's Gemini 2.5 Flash Image model), I've learned what separates good prompts from great ones.

Why Nano Banana Understands You Better​

Unlike traditional text-to-image models that rely on keyword matching, Nano Banana has deep language understanding. This means you can—and should—write prompts like you're describing a scene to another person, not just listing tags.

Instead of this:

cat, coffee shop, morning, cozy, warm lighting

Try this:

A photorealistic wide-angle shot of a tabby cat lounging on a windowsill 
in a cozy coffee shop. Morning sunlight streams through the window, 
creating warm, golden highlights on the cat's fur. The background shows 
blurred espresso machines and chalkboard menus.

The difference? Context and narrative. Nano Banana thrives on descriptive paragraphs.

The Six Essential Prompting Patterns​

1. Photorealistic Scenes: Think Like a Photographer​

For realistic images, use actual photography terminology. Mention camera angles, lens types, and lighting setups.

Template:

A [shot type] of [subject], [action or expression], set in [environment]. 
The scene is illuminated by [lighting description], creating a [mood] 
atmosphere. Captured with a [camera/lens details], emphasizing 
[key textures and details].

Real Example:

A close-up portrait of an elderly Japanese ceramicist carefully shaping 
a clay bowl on a pottery wheel. Natural window light from the left creates 
dramatic shadows, emphasizing the texture of the clay and the craftsman's 
weathered hands. Shot with an 85mm lens at f/2.8, shallow depth of field.

2. Product Photography: Studio-Quality Results​

Perfect for e-commerce or marketing materials without hiring a photographer.

Template:

A high-resolution, studio-lit product photograph of [product description] 
on a [background surface]. The lighting is [lighting setup] to highlight 
[specific feature]. The camera angle is [angle type]. Ultra-realistic, 
with sharp focus on [key detail].

Pro tip: Specify the lighting setup (three-point lighting, softbox, ring light) for more controlled results.

3. Illustrations and Stickers: Be Explicit About Style​

For icons, stickers, or design assets, clarity is everything.

Template:

A [style] sticker of a [subject], featuring [key characteristics] and 
a [color palette]. The design should have [line style] and [shading style]. 
The background must be transparent.

Example:

A kawaii-style sticker of a happy red panda holding a coffee cup, 
featuring big sparkly eyes and a cute smile. Use a pastel color palette 
with pink and orange tones. The design should have bold black outlines 
and flat cel shading. The background must be transparent.

4. Text in Images: Nano Banana's Superpower​

One of Nano Banana's standout features is accurate text rendering—something most AI models struggle with.

Template:

Create a [image type] with the text "[exact text]" in a [font style]. 
The design should be [style description], with a [color scheme].

Example:

Create a modern, minimalist logo for a coffee shop called 'The Daily Grind' 
with the text in a bold, geometric sans-serif font. The design should feature 
a stylized coffee cup integrated into the 'D', with a warm brown and cream 
color scheme.

5. Minimalist Compositions: Perfect for Presentations​

Great for creating backgrounds where text will be overlaid.

Template:

A minimalist composition featuring [subject] positioned in the 
[placement] of the frame. The background is a vast, empty [color] 
canvas, creating significant negative space. Soft, subtle lighting.

6. Sequential Art: For Visual Storytelling​

Template:

A single comic book panel in a [art style] style. In the foreground, 
[character description and action]. In the background, [setting details]. 
The panel has a dialogue box with the text "[Text]". The lighting creates 
a [mood] mood.

Advanced Techniques​

Iterative Refinement​

Don't expect perfection on the first try. Generate multiple variations and refine your prompts based on the results. If the first generation is close but needs adjustments, modify your prompt to be more specific about what you want to change.

Step-by-Step Instructions for Complex Scenes​

Break complex prompts into sequential steps:

First, create a background of a serene, misty forest at dawn. 
Then, in the foreground, add a moss-covered ancient stone altar. 
Finally, place a single, glowing sword on top of the altar.

Semantic Negative Prompts​

Instead of saying what you don't want, describe what you do want positively:

❌ "a street with no cars" ✅ "an empty, deserted street with parked bicycles along the sidewalk"

Control the Camera​

Use cinematic language to control composition:

  • "wide-angle shot" for expansive scenes
  • "macro shot" for extreme close-ups
  • "low-angle perspective" to make subjects look imposing
  • "bird's eye view" for overhead shots

Working with Context Images​

Nano Banana can use reference images alongside your text prompts. This is perfect for:

Style Transfer:

Transform this photograph of my living room into the artistic style 
of Van Gogh's Starry Night. Preserve the room's layout and furniture 
placement, but render it with bold, swirling brushstrokes and a vibrant 
blue and yellow palette.

Adding Elements:

Using the provided image of my product, place it on a modern marble 
countertop in a bright, minimalist kitchen. Ensure the product's features 
and branding remain completely unchanged.

Inpainting:

Using the provided image, change only the wall color to a sage green. 
Keep everything else exactly the same, preserving the original lighting 
and composition.

Making This Easier with Cubby Image​

If you're working in Photoshop, you can use Nano Banana directly without switching between apps. That's exactly why we built Cubby Image—to integrate AI image generation seamlessly into your Photoshop workflow.

With Cubby Image, you can:

  • Access Nano Banana (and other models) right from your Photoshop panels
  • Use your active selections as masks automatically
  • Pull context images directly from your layers
  • Generate multiple variations side-by-side without leaving your document

The key is having your tools work with your creative flow, not against it. When you're deep in a project, the last thing you want is to export files, switch to a browser, generate images, then import them back. Cubby Image keeps everything in one place.

Best Practices Checklist​

✅ Be hyper-specific - More detail = more control ✅ Provide context - Explain the image's purpose ✅ Use narrative descriptions - Not just keyword lists ✅ Iterate and refine - Generate multiple variations with adjusted prompts ✅ Think photographically - Use real camera/lighting terms ✅ Break down complexity - Step-by-step for complex scenes

Common Pitfalls to Avoid​

❌ Listing disconnected keywords ❌ Being vague about style or mood ❌ Forgetting to specify aspect ratio ❌ Not mentioning lighting for realistic scenes ❌ Using negative language ("no cars") instead of positive ("empty street")

Real-World Applications​

For Designers​

  • Product mockups for client presentations
  • Quick concept visualization
  • Marketing material backgrounds
  • Logo variations and iterations

For Content Creators​

  • Social media graphics
  • Blog post hero images
  • Thumbnail designs
  • Branded illustrations

For Photographers​

  • Mood board creation
  • Pre-visualization of complex shots
  • Background replacement concepts
  • Style reference generation

The Bottom Line​

Nano Banana's strength lies in its ability to understand natural language and detailed descriptions. The more specific and descriptive your prompts, the better your results will be.

Start with descriptive, narrative prompts. Use photography and artistic terminology. Generate multiple variations and refine your prompts based on results. Most importantly, experiment—every prompt teaches you something new about how to communicate with AI.

What has your experience been with AI image generation? Found any prompting techniques that work particularly well? I'd love to hear about them.

Try These Techniques in Your Workflow​

Ready to use Nano Banana in Photoshop? We built Cubby Image specifically to bring these AI models directly into your creative environment. No more switching between apps, exporting files, or losing context.

With Cubby Image, you get:

  • Nano Banana (plus FLUX, Qwen, and more) right in Photoshop
  • Automatic context from your selections and layers
  • Quick iteration by generating variations without re-exporting
  • Side-by-side comparison of variations
  • One-click upscaling with AI upscalers

The prompting techniques you learned here work even better when you can iterate quickly without breaking your creative flow.

Get Cubby Image → | Read the Docs

Questions about prompting or AI image generation? Feel free to reach out—I'm always happy to help troubleshoot or share what I've learned.