Image Gen Server

🚀 Image Generation Server

Image-Gen-Server is an image generation service based on Jimeng AI, designed to integrate with Cursor IDE. It can generate images according to text descriptions and provide download and save functions.

🚀 Quick Start

Image-Gen-Server is a powerful image generation service. It can generate high - quality images based on text descriptions and is seamlessly integrated with popular tools. You can follow the installation and integration steps below to start using it.

✨ Features

1. Perfect Integration: Supports seamless integration with tools such as Cursor and Claude.
2. Text - to - Image Generation: Generates high - quality images by inputting text descriptions.
3. Auto - Saving: Automatically generates and saves image files.
4. Custom Path: Allows setting the directory for saving images.
5. Batch Generation: Generates four images in different styles at once.

📦 Installation

1. Install using Smithery CLI

smith install image-gen-server

2. Manual Installation

1. Clone the project repository:

   git clone https://github.com/your-username/image-gen-server.git
   cd image-gen-server
   

2. Install dependencies:
   • Python Environment Requirement: Use Python 3.8 or higher.
   • Install Dependencies:

     pip install fastmcp proxy[jimeng]
     

3. Configure the project:
   • Modify the IMG_SAVA_FOLDER variable in the server.py file to set the default save path.
   • Configure the config.json file and add the API key of Jimeng AI.

🚀 Integration in Cursor

1. Open Cursor Settings

• Click Cursor > Settings in the menu bar.

2. Add an MCP Server

• Go to the Tools section and select MCP Servers.
• Click the Add Server button.
• Fill in the following information:
  • Name: Image-Gen-Server
  • Command:

    uv run --with fastmcp fastmcp run D:/path/to/server.py
    

  • Working Directory: Set it to the directory where the server script is located.

3. Complete the Configuration

• Click Add to complete the server addition.

💻 Usage Examples

Basic Usage

1. In Cursor's Agent mode:

   • Example 1: Generate a product logo according to project requirements and save it to the images folder in the project directory.
   • Example 2: Create a banner image for the website's home page.

2. Call the interface:

   from image_gen_server import generate_image
   
   images = generate_image(
       prompt="设计一个现代简约风格的logo",
       save_path="path/to/images/logo.jpg",
       width=1024,
       height=1024
   )
   

Advanced Usage

# In Cursor's Agent mode:
# Example 1: Generate a product logo
根据提供过你的项目需求，帮我生成一张产品logo，放在项目目录images下面

# Example 2: Design a website home - page banner
根据项目需求，帮我制作网站的首页，头部需要有banner图片。

🔧 Technical Details

• The server.py uses the fastmcp library to implement the MCP server function, providing services for Cursor and Claude.
• It calls the proxy.jimeng module to interact with Jimeng AI, supporting the following functions:
  • Image generation (generate_images)
  • Synchronous dialogue completion (create_completion)
  • Streaming dialogue completion (create_completion_stream)
  • Supports multi - account tokens
  • A complete error handling mechanism

📄 License

This project is licensed under the MIT License, and the author is Ling Feng.

💡 Troubleshooting

1. A black window pops up and disappears quickly after configuration

   • Possible reasons:
     • The server command is filled in incorrectly. Check the path for correctness, ensure it does not contain Chinese characters, and use a consistent path separator.
     • The environment is not installed correctly. Confirm that all dependencies have been installed.
     • The terminal used does not match the configuration. Try running it in different terminals (e.g., Git Bash, Cmd, PowerShell, or WSL).

2. Debug the server

   uv run --with fastmcp fastmcp dev D:/path/to/server.py
   

   This will open a debugging interface: http://localhost:5173/, and you can access the MCP Inspector through a browser for debugging.

For more detailed information, please refer to the project documentation and relevant technical materials.