Image Generation MCP Server

Generate images using Google's Gemini model through an MCP server.

Setup

1. Clone this repository

git clone https://github.com/MalluBeast69/gemini-img-gen-MCP
cd gemini-img-gen-MCP

2. Initialize a new MCP server project

uv init mcp-server-demo
cd mcp-server-demo

3. Install dependencies using uv

uv add "mcp[cli]"
uv add "google-generativeai"
uv add "Pillow"

4. Create an .mcp-config.json file in your workspace with the following content:

{
    "mcpServers": {
        "content_gen": {
            "command": "mcp",
            "args": ["run", "<PATH_TO_SERVER_PY>"],
            "env": {
                "GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
                "OUTPUT_IMAGE_PATH": "<YOUR_OUTPUT_PATH>"
            },
            "transportType": "stdio",
            "autoApprove": []
        }
    }
}

Replace:

• <PATH_TO_SERVER_PY> with the full path to your server.py file (e.g., "C:/Projects/image_gen/server.py")
• <YOUR_GEMINI_API_KEY> with your Gemini API key from https://makersuite.google.com/app/apikey
• <YOUR_OUTPUT_PATH> with where you want the generated images to be saved (e.g., "C:/Users/YourName/Pictures/GeneratedImages")

Usage

1. Start the MCP server:

mcp start content_gen

2. The server will now be ready to generate images based on your prompts!

Requirements

• Python 3.8+
• uv package manager
• Gemini API key