Nano Banana MCP Server

An MCP (Model Context Protocol) server for Google's Nano Banana Pro (Gemini 3 Pro Image) and Nano Banana (Gemini 2.5 Flash Image) AI image generation models.

Features

• 🎨 Text-to-Image Generation - Create images from natural language descriptions
• ✏️ Image Editing - Modify existing images with text prompts
• 🖼️ Multi-Image Composition - Combine up to 14 reference images
• 📐 Flexible Aspect Ratios - 10 aspect ratio options (1:1, 16:9, 9:16, etc.)
• 🔍 High Resolution - Up to 4K output with Nano Banana Pro
• 🔤 Text Rendering - Generate images with accurate, legible text
• 🌐 Google Search Grounding - Real-time information for current events/weather

Prerequisites

• Node.js 18 or higher
• Google AI API key from Google AI Studio

Installation

From Source

# Clone or download this directory
cd nanobanana-mcp-server

# Install dependencies
npm install

# Build the TypeScript
npm run build

Configuration

Set your API key as an environment variable:

# Linux/macOS
export GEMINI_API_KEY="your-api-key-here"

# Windows (PowerShell)
$env:GEMINI_API_KEY="your-api-key-here"

# Windows (CMD)
set GEMINI_API_KEY=your-api-key-here

Usage

As stdio Server (for Claude Desktop, etc.)

npm start

Or directly:

node dist/index.js

As HTTP Server

TRANSPORT=http PORT=3000 npm start

The server will listen at http://localhost:3000/mcp.

Claude Desktop Configuration

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "nanobanana": {
      "command": "node",
      "args": ["/path/to/nanobanana-mcp-server/dist/index.js"],
      "env": {
        "GEMINI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Available Tools

nanobanana_generate_image

Generate images from text descriptions.

Parameters:

Name	Type	Required	Description
prompt	string	Yes	Text description of the image
model	string	No	Model ID (default: gemini-3-pro-image-preview)
aspect_ratio	string	No	Aspect ratio (e.g., 1:1, 16:9)
resolution	string	No	Resolution: 1K, 2K, 4K (Pro only)
use_google_search	boolean	No	Enable real-time info grounding (Pro only)

Example:

{
  "prompt": "A photorealistic portrait of an astronaut on Mars at sunset, dramatic lighting, 85mm lens",
  "model": "gemini-3-pro-image-preview",
  "aspect_ratio": "16:9",
  "resolution": "4K"
}

nanobanana_edit_image

Edit an existing image using text prompts.

Parameters:

Name	Type	Required	Description
prompt	string	Yes	Description of the edit
image_base64	string	Yes	Base64-encoded image data
image_mime_type	string	Yes	MIME type (e.g., image/png)
model	string	No	Model ID
aspect_ratio	string	No	Output aspect ratio
resolution	string	No	Output resolution (Pro only)

Example:

{
  "prompt": "Add a wizard hat to the cat",
  "image_base64": "<base64-image-data>",
  "image_mime_type": "image/png"
}

nanobanana_compose_images

Compose images using multiple reference images (Nano Banana Pro only).

Parameters:

Name	Type	Required	Description
prompt	string	Yes	Description of composition
images	array	Yes	Array of {base64, mime_type} objects (max 14)
aspect_ratio	string	No	Output aspect ratio
resolution	string	No	Output resolution

Example:

{
  "prompt": "Create a group photo of these 5 people at a beach party",
  "images": [
    {"base64": "<image1>", "mime_type": "image/jpeg"},
    {"base64": "<image2>", "mime_type": "image/jpeg"}
  ],
  "aspect_ratio": "16:9",
  "resolution": "2K"
}

nanobanana_list_models

List available models and their capabilities.

Parameters:

Name	Type	Required	Description
response_format	string	No	markdown (default) or json

Models

Nano Banana Pro (gemini-3-pro-image-preview)

Google's most advanced image generation model. Best for:

• Professional asset production
• Images with accurate text rendering
• Complex compositions
• High-resolution output (up to 4K)
• Real-time information grounding

Nano Banana (gemini-2.5-flash-image)

Fast, low-latency image generation. Best for:

• Quick experimentation
• Batch processing
• Iterative design
• Cost-effective generation

Aspect Ratios

Ratio	Use Case
1:1	Square, social media posts
16:9	Landscape, presentations, YouTube thumbnails
9:16	Portrait, mobile screens, Stories
4:3	Classic photo format
3:2	DSLR photo format
21:9	Ultra-wide, cinematic

Tips for Best Results

Prompting

1. Be Descriptive: Include details about style, lighting, composition, colors, and mood
2. Use Photography Terms: For photorealistic images, mention camera angles, lens types, lighting setups
3. Specify Style: "Oil painting", "3D render", "watercolor", "photorealistic", etc.
4. Include Context: Describe the environment, atmosphere, and mood

Example Prompts

Photorealistic:

A photorealistic close-up portrait of an elderly Japanese ceramicist with 
deep wrinkles and a warm smile, inspecting a glazed tea bowl. Soft golden 
hour light from a window. Shot with 85mm portrait lens, bokeh background.

Stylized:

A kawaii-style sticker of a happy red panda wearing a bamboo hat, munching 
on a leaf. Bold outlines, cel-shading, vibrant colors. White background.

Logo:

Create a modern, minimalist logo for a coffee shop called 'The Daily Grind'. 
Clean, bold sans-serif font. Black and white. Circular design with a clever 
coffee bean element.

Error Handling

Common errors and solutions:

Error	Solution
"GEMINI_API_KEY is required"	Set the environment variable
Rate limit exceeded	Wait and retry, or reduce request frequency
Content policy violation	Modify prompt to comply with Google's policies
Invalid image format	Use supported formats: PNG, JPEG, GIF, WebP

Development

# Install dependencies
npm install

# Build TypeScript
npm run build

# Watch mode for development
npm run dev

License

MIT

Links

• Google AI Studio - Get your API key
• Gemini API Docs - Official documentation
• MCP Protocol - Learn about MCP