Image Gen With MCP

🚀 AI Image Generation Server Based on MCP Interface

This project offers a Model Context Protocol (MCP) server integrated with Stable Diffusion image generation capabilities, enabling AI agents to request and receive generated images. This project is based on the example provided in Block's Goose Custom Extension Tutorial.

🚀 Quick Start

📦 Installation

1. Create a virtual environment, force using .venv:

   virtualenv .venv
   

2. Activate the virtual environment:

   source .venv/bin/activate
   

3. Install the MCP package (for Goose integration):

   pip install -e .
   

💻 Starting the Service

The MCP server includes an integrated image generation service. You can start both with a single command:

Standard Mode:

source .venv/bin/activate  # Activate your virtual environment
export IMAGE_GEN_DIR=/absolute/path/to/folder # Set the target folder for generated images
image-gen-mcp

Development Mode (with FastMCP Inspector): Open two terminals:

Terminal 1:

source .venv/bin/activate  # Activate your virtual environment
export IMAGE_GEN_DIR=/absolute/path/to/folder # Set the target folder for generated images
image-gen-mcp # Start the image generation service (and an MCP server we won't use)

Terminal 2:

source .venv/bin/activate  # Activate your virtual environment
export IMAGE_GEN_DIR=/absolute/path/to/folder # Set the target folder for generated images
mcp dev src/image_gen_mcp/server.py # Start the MCP server with Inspector

Note: Only in development mode, the image generation service must be started separately.

This will start the MCP server with FastMCP Inspector, providing:

1. A web interface at http://127.0.0.1:8000 for debugging
2. The ability to check and adjust image generation settings

💻 Usage Examples

Direct API Access

Sending a Request: Use the curl command or any HTTP client to send a request to the service, for example:

curl -X POST "http://localhost:5000/api/generate" \
     --data-raw '{"prompt": "A flying owl", "negative_prompt": ""}'

Receiving a Response:

{
  "status": "success",
  "message": "Generation successful",
  "data": {
    "url": "http://localhost:5000/images/... png",
    "dimensions": "1024x1024"
  }
}

📚 File Organization

The project file structure is as follows:

.
├── src/
│   ├── image_gen_mcp/
│   │   ├── generator.py       # Main logic for handling image generation
│   │   └── server.py          # Server providing the MCP interface
│   └── requirements.txt      # Project dependencies
└── README.md                  # Project documentation

📚 Integration with Goose

Environment Variable Configuration: Add an environment variable IMAGE_GEN_DIR in Goose, specifying the location to store generated images.

Usage: In Goose, you can request to generate an image and receive an MCP object containing the URL and metadata. This service supports local operation, but it is recommended to use OpenAI GPT - 4o for faster response times.

🔧 Technical Details

Service Architecture

1. Image Generation Service (src/image_gen_mcp/generator.py)

   • Uses the Stable Diffusion model for actual image generation
   • Provides a simple HTTP API interface
   • Returns the image URL, dimensions, and metadata
   • Runs on port 5000 by default (can be customized)
   • Runs as an independent thread, coexisting with the MCP server

2. MCP Server (src/image_gen_mcp/server.py)

   • Provides a standardized MCP interface for AI agents to use
   • Forwards requests to the integrated image generation service
   • Returns an MCP object containing the URL and metadata

Stopping the Service

Press Ctrl+C to stop both services, as they are now running in the same process.