6 min read

Learn the basics of prompting Gemini AI Photo with straight to the point Prompts – from image generation to photo analysis

Ever stared at a blank screen, a wacky idea buzzing in your brain—a duck wearing a detective trench coat, a floating bookstore in the clouds, your dog as a medieval knight—but your artistic skills topped out at stick figures? You’re not alone. We’ve all been there, stranded in the chasm between wild imagination and tangible creation. But what if I told you that gap is now a tiny crack you can step over with a few perfectly chosen words? Welcome to the new era of visual storytelling, where creating stunning AI-generated images for beginners is less about complex software and more about learning the secret language of prompts.

This isn’t just about asking a robot for a picture. This is about becoming a digital art director from your smartphone, conjuring photorealistic landscapes, whimsical character designs, and brand-ready graphics before your coffee gets cold. And the best part? The tools are already in your pocket, from the cloud-powered genius of advanced models to the on-device magic of Gemini Nano Banana for quick photo analysis that can kickstart your entire creative process.

Forget years of art school. The new canvas is a text box. The new paintbrush is your vocabulary. In this guide, we’re cracking open the playbook on simple prompt engineering techniques for unique AI art, serving up the exact phrases and formulas that transform “I wish I could see that” into “Look what I made!” We’ll cover everything from generating specific artistic styles with AI image prompts to using your camera with Gemini Nano for instant visual brainstorming and object identification.

Examples for Gemini Nano “Banana” (On-Device, Lightweight)

Gemini Nano is designed for fast, on-device tasks with simple prompts. It’s great for quick photo analysis without needing an internet connection.

Example 1: Basic Object & Scene Recognition

• Your Prompt: What is this?
• You Do: Take a photo of your pet or a plant.
• Nano Might Respond: “That’s a golden retriever dog.” or “That’s a potted snake plant.”

Example 2: Simple Description & Context

• Your Prompt: Describe this scene.
• You Do: Take a photo of your desk or a park bench.
• Nano Might Respond: “A cluttered desk with a laptop, coffee mug, and notebook.” or “A sunny park with an empty bench and trees.”

Example 3: Extracting Clear Text (OCR)

• Your Prompt: Read the text.
• You Do: Take a photo of a product label or a street sign.
• Nano Might Respond: It will transcribe the visible text from the image.

Example 4: Basic Comparison

• Your Prompt: Are these two objects the same?
• You Do: Take photos of two similar shoes or fruits.
• Nano Might Respond: “Yes, both are red apples.” or “No, one is a sneaker and the other is a boot.”

Key Difference for Nano “Banana”:

• Keep it simple. Nano is for quick, straightforward recognition and description, not complex creative storytelling or extensive analysis.
• It works offline. Perfect for quick checks when you have no signal.
• Prompts are often single actions: “What is this?”, “Describe.”, “Read this.”

Pro-Tip:

Always remember to attach or allow camera access for the photo you want Gemini to analyze. Start your prompt after you’ve shared the image.

Good Nano Prompt: (Attach photo of receipt) "What was the total cost?"
Good Advanced Prompt: (Attach photo of a painting) "Explain the artistic style and suggest two similar artists."

Choose the model based on your need: Nano for quick, on-device facts; Advanced/Pro for detailed analysis and creative tasks.

For Direct Image Generation (Using Imagen/Image FX)

When you prompt Gemini to create an image, it typically activates Google’s image generation model in the background.

1. Basic Object & Scene Creation

• “Create a photo of a fluffy cat sleeping in a sunbeam on a wooden floor.”
• “Generate an image of a futuristic city at night with flying cars.”
• “Show me a watercolor painting of a mountain lake at sunrise.”

2. Style & Artistic Direction

• “Pixel art of a wizard with a blue hat.”
• “A minimalist logo for a coffee shop called ‘Bean There’.”
• “An oil painting in the style of Van Gogh of a starry night over a modern suburb.”
• “A 1950s sci-fi magazine cover about robots.”

3. For Specific Formats & Uses

• “A square Instagram post about beach vacation vibes, with text that says ‘Paradise Found’.”
• “A wide banner image for a website header showing a diverse team collaborating.”
• “A simple icon of a cloud with an arrow pointing down for a download button.”

4. Creative & Whimsical Concepts

• “A teddy bear astronaut planting a flag on a cookie planet.”
• “A steampunk version of a hummingbird.”
• “An enchanted forest where the trees have crystal leaves.”

---

For Gemini Nano “Banana” (On-Device)

Important: Gemini Nano does NOT create or generate new images. It can only analyze, describe, and answer questions about photos you give it.

What Nano CAN do with your photos:

• Generate descriptive text from an image.
• Answer questions about image content.
• Extract and summarize text from photos.

Nano Prompt Example (Analysis, not Creation):

1. Take a photo of your sketch.
2. Ask Nano: "Describe this drawing in detail."
3. Nano will give you a text description (e.g., “A rough pencil sketch of a house with two windows and a smoking chimney.”).

---

Pro-Tips for Better Image Creation

1. Be Specific: Add details about style, composition, color, mood, and perspective.
   • Less specific: “A dog in a park.”
   • More specific: “A photorealistic image of a happy golden retriever running through a green park in spring, low-angle shot, with sun flares.”
2. Use Style References: Mention art styles, mediums, or even photographer names.
   • “Ansel Adams style black and white landscape of Yosemite.”
   • “Studio Ghibli animated style scene of a friendly forest spirit.”
3. Combine Concepts: Mash up ideas for something unique.
   • “A cyberpunk sushi restaurant in a rainy alley.”
   • “A Renaissance painting of a scientist looking through a microscope.”
4. Iterate: Use the first result as a starting point.
   • “Now make that cat orange and white.”
   • “Make the scene at dusk instead of noon.”

To create images in the Gemini ecosystem:

1. Go to the Gemini web/app interface or labs.google.
2. Type a creation prompt (like the ones above).
3. Gemini will either generate the image directly (via integrated Image FX/Imagen) or guide you to the correct tool.

Example Flow:
You: “Create a logo for my dog walking business.”
Gemini/Image FX: Generates several logo options with dog-related imagery.