README.md•1.85 kB
# OpenAI Image Generation MCP Server
This MCP (Model Context Protocol) server provides an interface to generate images using OpenAI's DALL-E image generation model.
## Setup
1. Install dependencies:
```
npm install
```
2. Create a `.env` file in the root directory with your OpenAI API key:
```
OPENAI_API_KEY=your-api-key-here
```
3. Build the project:
```
npm run build
```
4. Make the server executable:
```
chmod +x build/index.js
```
5. Link the package (optional, for global access):
```
npm link
```
## Image Storage
Generated images are automatically saved to an `image` directory in the server's working directory. Each image is saved with a filename that includes:
- Timestamp
- Simplified version of the prompt
- Image sequence number
For example: `2023-04-26T12-45-32-000Z_cat_wearing_a_space_suit_1.png`
## Usage with Claude Desktop
1. Add this server to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"openai-image": {
"command": "openai-image-server",
"env": {
"OPENAI_API_KEY": "your-api-key"
}
}
}
}
```
2. Restart Claude Desktop and look for the OpenAI image server in the 🔌 menu.
## Image Generation Tool
The server provides a single tool:
- **generate_image**: Generates images using OpenAI's DALL-E model
### Parameters:
- `prompt` (required): The text prompt to generate an image from
- `n` (optional): Number of images to generate (1-10, default: 1)
- `size` (optional): Image size (256x256, 512x512, 1024x1024, 1792x1024, or 1024x1792, default: 1024x1024)
- `model` (optional): The model to use (default: "gpt-image-1")
## Example
When connected to Claude Desktop, you can generate images with prompts like:
"Generate an image of a cat wearing a space suit on the moon"