Universal Image Generator MCP

🚀 Universal Image Generator MCP Server

This project is a multi - provider AI image generation server for MCP clients. It allows users to generate high - quality images using Google (Imagen & Gemini), ZHIPU AI CogView - 4, or Alibaba Bailian through any MCP - compatible application. It's a fork and rewrite of the original: https://github.com/qhdrl12/mcp-server-gemini-image-generator repo.

✨ Features

• Multi - Provider Support: You can choose between Google (Imagen/Gemini), ZhipuAI, or Bailian.
• Image Generation: All providers support text - to - image generation.
• Image Transformation: Edit existing images (only available for Google & Bailian).
• Smart Language Optimization: Automatically translate and optimize prompts.
• Local Storage: Save generated images to your specified directory.

🚀 Quick Start

📦 1. Install via uvx

No manual installation required! The server will be automatically downloaded and run.

🔑 2. Get API Keys

Choose one provider and obtain an API key:

• Google: Google AI Studio (supports both Imagen and Gemini models)
• ZhipuAI: ZHIPU AI Platform
• Bailian: Alibaba DashScope

⚙️ 3. Configure MCP Client

Add the following to your MCP client configuration (e.g., claude_desktop_config.json):

{
    "mcpServers": {
        "universal-image-generator": {
            "command": "uvx",
            "args": [
                "universal-image-generator-mcp"
            ],
            "env": {
                "IMAGE_PROVIDER": "google",
                "GOOGLE_MODEL": "gemini",
                "ZHIPU_API_KEY": "your-api-key-here",
                "GEMINI_API_KEY": "your-api-key-here",
                "DASHSCOPE_API_KEY": "your-api-key-here",
                "OUTPUT_IMAGE_PATH": "/path/to/save/images"
            }
        }
    }
}

Environment Variables:

• IMAGE_PROVIDER: "google", "zhipuai", or "bailian"
• GOOGLE_MODEL: "gemini" or "imagen" (only for Google provider, defaults to "gemini")
• Set the corresponding API key for your chosen provider
• OUTPUT_IMAGE_PATH: Directory to save generated images (optional)

💻 Usage Examples

📝 Basic Usage

Once configured, you can ask your AI assistant:

• "Generate an image of a sunset over mountains"
• "Create a 3D rendered flying pig in a sci - fi city"
• "Transform this image by adding snow to the scene"

Generated images will be saved to your configured output directory.

🖼️ Example Output

Prompt: "Create a 3D rendered image of a pig with wings and a top hat flying over a futuristic sci - fi city with lots of greenery"

Transform: "Add a cute baby whale flying alongside the pig"

📚 Documentation

🛠️ Available Tools

generate_image_from_text

Create images from text descriptions.

generate_image_from_text(prompt: str, model_type: Optional[str] = None) -> str

Parameters:

• prompt: Text description of the image to generate
• model_type: Optional model selection for Google provider ("gemini" or "imagen"). Only applies to Google provider. If not specified, it uses the GOOGLE_MODEL environment variable (defaults to "gemini").

transform_image_from_encoded (Google & Bailian only)

Transform images using base64 - encoded image data.

transform_image_from_encoded(encoded_image: str, prompt: str) -> str

transform_image_from_file (Google & Bailian only)

Transform existing image files.

transform_image_from_file(image_file_path: str, prompt: str) -> str

🔧 Technical Details

📊 Provider Capabilities

Note: For the Google provider, image transformation is only supported with Gemini models. Imagen is for generation only.

🛠️ Development

Test the server locally:

git clone https://github.com/ECNU3D/universal-image-generator-mcp.git
cd universal-image-generator-mcp
fastmcp dev src/universal_image_generator_mcp/server.py

Visit http://localhost:5173/ to use the MCP Inspector for testing.

📄 License

This project is under the MIT License.