MCP Nano Banana

🚀 MCP Nano Banana

This project is an MCP (Model Context Protocol) server that generates images using the Google Gemini API.

🚀 Quick Start

To use this server, you need a Google Gemini API key and an ImgBB API key. Access https://api.imgbb.com/ to generate an IMGBB API Key, which is used to store and host the image online.

✨ Features

This server implements the Model Context Protocol to expose a single tool, generate_image, to a compatible AI model. The tool accepts a text prompt, uses the Google Gemini API to generate an image, saves the image to the public/ directory for auditing, and returns the raw image data as a base64-encoded string.

📦 Installation

1. Dependencies

This project uses Python and its dependencies are defined in pyproject.toml. You can install them using pip:

pip install .
# Or
uv sync

This will install mcp, google-generativeai, and other required packages.

2. API Key

1. Create a file named .env in the root of the project.
2. Add your API key to the .env file in the following format:

GEMINI_API_KEY="YOUR_API_KEY_HERE"
IMGBB_API_KEY="YOUR_API_KEY_HERE"

💻 Usage Examples

To use the server with Claude Desktop or other applications

{
  "mcpServers": {
    "mcp-nano-banana": {
        "command": "uvx",
        "args": [
            "mcp-nano-banana"
        ],
        "env": {
            "GEMINI_API_KEY": "YOUR_API_KEY_HERE",
            "IMGBB_API_KEY": "YOUR_API_KEY_HERE"
        }
    }
  }
}

Running the Server

This server is designed to be run as a subprocess by an MCP client or using the mcp command-line tool. The server listens for requests on stdio.

uvx --from git+https://github.com/GuilhermeAumo/mcp-nano-banana mcp-nano-banana

📚 Documentation

Publishing new pipy version

To publish a new version of this package to PyPI:

1. Update the version
   Edit the version field in pyproject.toml to the new version number.

2. Build the package
   Run:

   uv build
   

   This will create .tar.gz and .whl files in the dist/ directory.

3. Upload to PyPI

   uv publish
   

4. Tag the release (optional but recommended)
   Commit the changes to github first, then:

   git tag v<new-version>
   git push --tags
   

⚠️ Important Note

You need a PyPI account and must be listed as a maintainer of the project.

For more details, see the Python Packaging User Guide.