DALL-E MCP Server

# DALL-E MCP Server Installation Guide Complete step-by-step installation guide for setting up the DALL-E MCP server with various AI clients. ## Prerequisites Before installing, ensure you have: - **Node.js** (version 18 or higher) - [Download from nodejs.org](https://nodejs.org/) - **OpenAI API Key** - [Get from OpenAI Platform](https://platform.openai.com/api-keys) - **MCP-compatible client** (Claude Desktop, ChatGPT, or custom implementation) ## Step 1: Download and Setup ### Option A: Clone from Repository ```bash git clone <repository-url> cd dall-e-mcp-server ``` ### Option B: Download and Extract 1. Download the project files 2. Extract to your desired directory 3. Open terminal in the project directory ## Step 2: Install Dependencies ```bash # Install all required packages npm install # Verify installation npm list ``` Expected output should show packages like: - `@modelcontextprotocol/sdk` - `openai` - `fs-extra` ## Step 3: Configure Environment ### Create Environment File ```bash # Copy the example environment file cp .env.example .env ``` ### Edit Environment Variables Open `.env` file and configure: ```env # Required: Your OpenAI API Key OPENAI_API_KEY=sk-your-actual-openai-api-key-here # Optional: Customize defaults DEFAULT_IMAGE_SIZE=1024x1024 DEFAULT_QUALITY=standard OUTPUT_DIRECTORY=./generated_images ``` **⚠️ Important:** Replace `sk-your-actual-openai-api-key-here` with your real OpenAI API key. ### Verify API Key Test your API key works: ```bash # Quick test (optional) node -e " const OpenAI = require('openai').default; const client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY }); console.log('API key configured correctly!'); " ``` ## Step 4: Build the Server ```bash # Compile TypeScript to JavaScript npm run build # Verify build succeeded ls dist/ ``` You should see `index.js` and related files in the `dist/` directory. ## Step 5: Test Installation ```bash # Quick test - server should start and show "running on stdio" timeout 3s npm start || echo "✅ Server test completed" # On Windows (PowerShell): # Start-Job { npm start } | Wait-Job -Timeout 3 | Remove-Job ``` Expected output: `DALL-E MCP Server running on stdio` --- ## Client Integration Choose your AI client and follow the appropriate setup: ### 🔷 Claude Desktop Integration #### macOS/Linux Setup 1. **Find config location:** ```bash # Config file location ~/.config/claude/claude_desktop_config.json ``` 2. **Create/edit config file:** ```bash # Create directory if it doesn't exist mkdir -p ~/.config/claude # Edit config file nano ~/.config/claude/claude_desktop_config.json ``` 3. **Add server configuration:** ```json { "mcpServers": { "dall-e-image-generator": { "command": "node", "args": ["/FULL/PATH/TO/dall-e-mcp-server/dist/index.js"], "env": { "OPENAI_API_KEY": "sk-your-actual-api-key-here" } } } } ``` 4. **Update the path:** ```bash # Get full path to your project pwd # Copy this path and replace /FULL/PATH/TO/dall-e-mcp-server above ``` #### Windows Setup 1. **Find config location:** ``` %APPDATA%\Claude\claude_desktop_config.json ``` 2. **Create/edit config file:** - Open File Explorer - Type `%APPDATA%\Claude` in address bar - Create `claude_desktop_config.json` if it doesn't exist 3. **Add server configuration:** ```json { "mcpServers": { "dall-e-image-generator": { "command": "node", "args": ["C:\\FULL\\PATH\\TO\\dall-e-mcp-server\\dist\\index.js"], "env": { "OPENAI_API_KEY": "sk-your-actual-api-key-here" } } } } ``` 4. **Get Windows path:** ```cmd # In your project directory cd # Copy this path and use double backslashes in config ``` #### Restart Claude Desktop - Completely quit Claude Desktop - Restart the application - The MCP server should now be available ### 🔷 ChatGPT Integration ChatGPT integration depends on your setup: #### OpenAI API with MCP Support ```json { "mcpServers": { "dall-e-generator": { "command": "node", "args": ["/path/to/dall-e-mcp-server/dist/index.js"], "env": { "OPENAI_API_KEY": "your-api-key" } } } } ``` #### Custom Implementation Use the OpenAI Agents SDK or Responses API to connect to the MCP server. ### 🔷 Other MCP Clients For custom MCP clients, connect via stdio transport: ```typescript import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js"; import { Client } from "@modelcontextprotocol/sdk/client/index.js"; const transport = new StdioClientTransport({ command: "node", args: ["/path/to/dall-e-mcp-server/dist/index.js"], env: { OPENAI_API_KEY: "your-api-key" } }); const client = new Client({ name: "dall-e-client", version: "1.0.0" }, {}); await client.connect(transport); ``` --- ## Step 6: Verify Installation ### Test with Claude Desktop 1. Open Claude Desktop 2. Start a new conversation 3. Try generating an image: ``` Can you generate an image of a sunset over mountains? ``` 4. Claude should use the `generate_image` tool and save the image locally ### Check Generated Images ```bash # Check if images are being saved ls generated_images/ # View a generated image (macOS) open generated_images/dalle_*.png # View a generated image (Linux) xdg-open generated_images/dalle_*.png # View a generated image (Windows) start generated_images\dalle_*.png ``` --- ## Troubleshooting ### Common Issues #### ❌ "OPENAI_API_KEY is missing" **Solution:** 1. Verify `.env` file exists and contains your API key 2. Check that the API key starts with `sk-` 3. Ensure no extra spaces around the key #### ❌ "Command not found: node" **Solution:** 1. Install Node.js from [nodejs.org](https://nodejs.org/) 2. Restart terminal after installation 3. Verify with: `node --version` #### ❌ "Cannot find module" **Solution:** ```bash # Reinstall dependencies rm -rf node_modules package-lock.json npm install npm run build ``` #### ❌ "Permission denied" **Solution:** ```bash # Make script executable (macOS/Linux) chmod +x dist/index.js # Or run with explicit node command node dist/index.js ``` #### ❌ Claude Desktop doesn't see the server **Solution:** 1. Check config file path is correct 2. Verify JSON syntax is valid 3. Use absolute paths, not relative paths 4. Restart Claude Desktop completely 5. Check Claude Desktop logs/console for errors #### ❌ "Failed to download image" **Solution:** 1. Check internet connection 2. Verify OpenAI API credits/billing 3. Try a simpler prompt first ### Debug Mode Enable detailed logging: ```bash # Set debug environment export DEBUG=dall-e-mcp:* # Run server with debug info npm run dev ``` ### Check Server Status ```bash # Test server manually echo '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | node dist/index.js ``` Expected response should include the `generate_image` tool. ### Validate Configuration Create a test script: ```bash # Create test file cat > test_config.js << 'EOF' const config = require('./claude_desktop_config.json'); console.log('Config loaded successfully:', config); EOF # Run test node test_config.js ``` --- ## Advanced Configuration ### Custom Output Directory ```env OUTPUT_DIRECTORY=/custom/path/to/images ``` ### Multiple API Keys (Enterprise) ```json { "mcpServers": { "dall-e-prod": { "command": "node", "args": ["/path/to/dall-e-mcp-server/dist/index.js"], "env": { "OPENAI_API_KEY": "prod-key-here" } }, "dall-e-dev": { "command": "node", "args": ["/path/to/dall-e-mcp-server/dist/index.js"], "env": { "OPENAI_API_KEY": "dev-key-here", "OUTPUT_DIRECTORY": "./dev_images" } } } } ``` ### System Service (Optional) For production deployments, create a system service: #### Linux (systemd) ```bash # Create service file sudo nano /etc/systemd/system/dall-e-mcp.service ``` ```ini [Unit] Description=DALL-E MCP Server After=network.target [Service] Type=simple User=your-username WorkingDirectory=/path/to/dall-e-mcp-server Environment=OPENAI_API_KEY=your-key ExecStart=/usr/bin/node dist/index.js Restart=always [Install] WantedBy=multi-user.target ``` ```bash # Enable and start service sudo systemctl enable dall-e-mcp sudo systemctl start dall-e-mcp ``` --- ## Security Notes 1. **Never commit `.env` files** - they contain your API key 2. **Restrict file permissions:** ```bash chmod 600 .env ``` 3. **Use environment variables** in production instead of `.env` files 4. **Monitor API usage** to prevent unexpected charges 5. **Rotate API keys** regularly --- ## Next Steps After successful installation: 1. **Read the [README.md](./README.md)** for usage examples 2. **Experiment with different prompts** and parameters 3. **Check the `generated_images/` directory** for your creations 4. **Monitor OpenAI API usage** in your dashboard 5. **Customize the server** for your specific needs --- ## Support If you encounter issues: 1. **Check this troubleshooting section** first 2. **Verify your OpenAI API key** and billing status 3. **Test with a minimal prompt** to isolate issues 4. **Check MCP client logs** for detailed error messages 5. **Ensure all prerequisites** are properly installed --- **🎉 Congratulations!** Your DALL-E MCP server should now be ready to generate amazing images through your AI client!