Content
[](https://mseep.ai/app/fengin-image-gen-server)
# Image-Gen-Server
<div align="center">
<img src="images/logo_0.png" alt="Image-Gen-Server Logo" width="100%">
</div>
[](https://smithery.ai/server/@fengin/image-gen-server)
Image-Gen-Server is an image generation service based on JiMeng AI, specifically designed for integration with Cursor IDE. It receives text descriptions from Cursor, generates corresponding images, and provides functionality for downloading and saving images.
You can view the development process of this plugin on my website: [Developing an MCP Server Integrated with Cursor, Giving Cursor Wings!](https://aibook.ren/archives/mcp-server-for-cursor)
For more AI knowledge, see the AI Book (https://aibook.ren).
<div align="center">
<img src="images/example.png" alt="Image-Gen-Server Example" width="100%">
</div>
## Features
- Perfect integration with Cursor IDE
- Supports text-to-image generation
- Automatically saves generated images
- Supports custom save paths
- Generates four images at once for more options
## Installation
### Installing via Smithery
To install Image-Gen-Server for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@fengin/image-gen-server):
```bash
npx -y @smithery/cli install @fengin/image-gen-server --client claude
```
1. Environment preparation: MCP is a relatively new technology, and the dependency environment versions are quite recent.
- python 3.10+
- Install npm
- Install nodejs (versions v15 and v16 have been tested and do not work; v20 has been verified in the development environment, others are unverified)
- Install pip install uv
- If you want to debug, you also need to install this: npm install -g @modelcontextprotocol/inspector@0.4.0
2. Clone the project
```bash
git clone https://github.com/fengin/image-gen-server.git
cd image-gen-server
```
3. Install dependencies
```bash
pip install -r requirements.txt
pip install uv
```
4. Set JiMeng Token and default image save path
Modify these two configurations in the server.py file
```bash
# API configuration
JIMENG_API_TOKEN = "057f7addf85dxxxxxxxxxxxxx" # Your session_id obtained from JiMeng, supports multiple, separated by commas
IMG_SAVA_FOLDER = "D:/code/image-gen-server/images" # Default image save path
```
## Cursor Integration
<div align="center">
<img src="images/cursor_config.png" alt="Cursor Configuration" width="100%">
</div>
1. Open Cursor settings
- Click the settings icon in the lower left corner
- Select Features > MCP Servers
- Click "Add new MCP server"
2. Fill in the server configuration
- Name: `image-gen-server` (or any name you prefer)
- Type: `command`
- Command:
```bash
uv run --with fastmcp fastmcp run D:\code\image-gen-service\server.py
```
Note: Replace the path with your actual project path
- Windows example: `uv run --with fastmcp fastmcp run D:/code/image-gen-service/server.py`
- macOS/Linux example: `uv run --with fastmcp fastmcp run /Users/username/code/image-gen-server/server.py`
There are many path issues on Windows, try various slashes for D:/code/image-gen-server/server.py.
After filling in, a black window will pop up, and then you can ask Cursor to generate the images you need. Currently, the black window will keep running, and there is no way to resolve this issue yet.
## Usage
In Cursor, to generate images, you need to prompt it to understand how to use the image tool in agent mode, and then directly state your image generation requirements and save location.
## Obtaining JiMeng Token
1. Visit [JiMeng](https://jimeng.jianying.com/)
2. Log in to your account
3. Press F12 to open the developer tools
4. Find `sessionid` in Application > Cookies
5. Set the found sessionid in server.py's JIMENG_API_TOKEN
## Function Description
### generate_image
```python
async def generate_image(prompt: str, file_name: str, save_folder: str = None, sample_strength: float = 0.5, width: int = 1024, height: int = 1024) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
"""Generate an image based on text description
Args:
prompt: The text prompt description for the image
file_name: The filename for the generated image (without path; defaults to .jpg if no suffix is provided)
save_folder: The absolute directory for saving the image (optional; defaults to IMG_SAVA_FOLDER)
sample_strength: The fineness of the generated image (optional; range 0-1; defaults to 0.5)
width: The width of the generated image (optional; defaults to 1024)
height: The height of the generated image (optional; defaults to 1024)
Returns:
List: A JSON string containing the generation results
"""
```
### Technical Implementation
1. server.py uses fastmcp to implement the capabilities of an MCP server, providing services for Cursor/Claude.
2. server.py calls the proxy.jimeng module to interact with JiMeng AI in reverse. The proxy.jimeng reverse module can also be installed separately and mainly provides the following key functionalities:
- Image generation (generate_images)
- Synchronous dialogue completion (create_completion)
- Streaming dialogue completion (create_completion_stream)
- Multi-account token support
- Complete error handling
For more detailed information, please refer to `proxy/jimeng/README.md`.
### Usage Examples
```cmd
# In Cursor agent mode
# Example 1
Based on the project requirements you provided, help me generate a product logo and place it in the project directory under images.
# Example 2
Based on the project requirements, help me create the homepage of the website, with a banner image at the top.
```
## License
MIT License
Author: Ling Feng
## Troubleshooting
1. After configuration, a black window pops up and quickly disappears, and the tool status changes to No tools found.
Reason: Did not start properly, usually due to the following reasons:
- Incorrect configuration command; check if the command is correct, usually the server.py path is incorrect, or the path contains Chinese characters, or the slashes are incorrect.
- The required environment is not prepared.
- The terminal for running dependencies is incorrect. For example, on Windows, there are git bash, cmd, powershell, wsl, etc. Try these terminals. The default terminal for my Cursor configuration is cmd. If you encounter errors running in the corresponding terminal, it is usually due to the environment not being set up correctly; installing the environment should resolve it.
2. After running normally, if you want to see the call logs or debug, how to do it?
Change the command to the following:
```
uv run --with fastmcp fastmcp dev D:/code/image-gen-service/server.py
```
Just change the last `run` to `dev`.
Alternatively, find a terminal and run the following command to enter debug mode:
```
fastmcp dev D:/code/image-gen-service/server.py
```
A debug address will be output: http://localhost:5173/, you can open this address in your browser to use the MCP Inspector for debugging. For specific usage of MCP Inspector, please refer to the official documentation.
