MCP Server for ElevenLabs

# 🎯 ElevenLabs MCP Server - Ready for Integration! Your custom MCP server is now ready to be integrated with your ElevenLabs conversational agent! ## ✅ What's Ready - **HTTP MCP Server**: Running on `http://localhost:3000/mcp` - **3 Custom Tools**: search_web, get_weather, calculate - **Production Ready**: Docker support, health checks, error handling - **ElevenLabs Compatible**: Follows MCP protocol 2024-11-05 ## 🚀 Quick Integration Steps ### 1. Start Your Server ```bash # Development mode npm run dev:elevenlabs # OR Production mode npm run build npm run start:elevenlabs # OR Docker docker-compose up -d ``` ### 2. Configure in ElevenLabs Dashboard 1. Go to [ElevenLabs Dashboard](https://elevenlabs.io/) 2. Navigate to **Conversational AI** → **Your Agent** 3. Click **Customization** tab 4. Under **MCP Servers**, click **Add Server** 5. Enter: `http://localhost:3000/mcp` 6. Save configuration ### 3. Test Your Integration Your agent can now use these tools: ``` User: "What's the weather in London?" Agent: [Uses get_weather tool] "The current weather in London is 18°C with partly cloudy skies." User: "Calculate 25% of 400" Agent: [Uses calculate tool] "25% of 400 is 100." User: "Search for the latest AI news" Agent: [Uses search_web tool] "Here are the latest AI developments..." ``` ## 🔧 Server Details - **URL**: `http://localhost:3000/mcp` - **Health Check**: `http://localhost:3000/health` - **Protocol**: HTTP/JSON-RPC 2.0 - **CORS**: Enabled for cross-origin requests - **Error Handling**: Comprehensive error responses ## 🛠️ Available Tools ### 1. search_web **Purpose**: Search the web for current information **Parameters**: - `query` (string): Search term - `maxResults` (number, optional): Number of results ### 2. get_weather **Purpose**: Get current weather for any location **Parameters**: - `location` (string): City/location name - `units` (string, optional): "celsius" or "fahrenheit" ### 3. calculate **Purpose**: Perform mathematical calculations **Parameters**: - `expression` (string): Math expression (e.g., "2 + 2 * 3") ## 🌐 Deployment Options ### Local Development ```bash npm run dev:elevenlabs ``` ### Production (PM2) ```bash npm install -g pm2 pm2 start dist/elevenlabs-server.js --name "elevenlabs-mcp" pm2 save pm2 startup ``` ### Docker ```bash docker-compose up -d ``` ### Cloud Deployment - **Railway**: Connect GitHub repo - **Render**: Deploy from Docker - **Heroku**: Use Procfile - **Vercel**: Serverless deployment ## 🔍 Testing ### Health Check ```bash curl http://localhost:3000/health ``` ### List Tools ```bash curl -X POST http://localhost:3000/mcp \ -H "Content-Type: application/json" \ -d '{"jsonrpc": "2.0", "id": 1, "method": "tools/list"}' ``` ### Test Tool Call ```bash curl -X POST http://localhost:3000/mcp \ -H "Content-Type: application/json" \ -d '{"jsonrpc": "2.0", "id": 2, "method": "tools/call", "params": {"name": "calculate", "arguments": {"expression": "2 + 2"}}}' ``` ## 📝 Adding Custom Tools 1. **Define Schema** in `src/tools.ts`: ```typescript export const MyToolSchema = z.object({ parameter: z.string().describe("Description"), }); ``` 2. **Implement Handler**: ```typescript export async function handleMyTool(args: z.infer<typeof MyToolSchema>) { // Your logic here return { content: [{ type: "text" as const, text: "Result" }], }; } ``` 3. **Add to Server** in `src/elevenlabs-server.ts`: ```typescript case "my_tool": result = await handleMyTool(args); break; ``` 4. **Rebuild & Restart**: ```bash npm run build npm run dev:elevenlabs ``` ## 🚨 Troubleshooting ### Server Won't Start - Check port availability: `netstat -an | findstr :3000` - Try different port: `PORT=3001 npm run dev:elevenlabs` ### ElevenLabs Can't Connect - Verify server is running: `curl http://localhost:3000/health` - Check firewall settings - Ensure URL format: `http://localhost:3000/mcp` ### Tools Not Working - Check server logs for errors - Verify tool names match exactly - Test with curl commands above ## 🔒 Security Notes - **Production**: Use HTTPS - **Authentication**: Add API keys if needed - **Rate Limiting**: Implement request throttling - **Input Validation**: All inputs validated with Zod ## 📞 Support - **MCP Server Issues**: Check this repository - **ElevenLabs Integration**: [ElevenLabs Support](https://elevenlabs.io/support) - **MCP Protocol**: [MCP Documentation](https://modelcontextprotocol.io/) --- ## 🎉 You're All Set! Your ElevenLabs conversational agent now has access to custom tools through your MCP server. The agent can search the web, get weather information, and perform calculations - making it much more capable and useful for your users! **Next Steps:** 1. Deploy to production 2. Add more custom tools 3. Monitor usage and performance 4. Iterate and improve based on user feedback