Brummer MCP Server

--- sidebar_position: 1 --- # MCP Integration Overview Brummer includes a built-in Model Context Protocol (MCP) server that allows external tools like VSCode, Claude Code, and Cursor to interact with your development processes. ## What is MCP? The Model Context Protocol (MCP) is an open standard that enables seamless communication between development tools and services. It provides a unified interface for: - **Process Management** - Start, stop, and monitor processes - **Log Access** - Stream and search logs in real-time - **Error Detection** - Get notified of errors and warnings - **System Status** - Monitor resource usage and performance ## Why Use MCP with Brummer? ### For AI Assistants AI coding assistants can: - **Understand your dev environment** - See what's running and its status - **Debug issues** - Access error logs and stack traces - **Execute commands** - Run scripts and manage processes - **Monitor performance** - Track memory and CPU usage ### For IDEs and Editors Development environments can: - **Integrate process management** - Control Brummer from your editor - **Display inline errors** - Show errors next to your code - **Quick navigation** - Jump to error locations - **Real-time feedback** - See build and test results instantly ### For Custom Tools Build your own integrations: - **Automation scripts** - Programmatically control processes - **Monitoring dashboards** - Create custom status displays - **CI/CD pipelines** - Integrate with build systems - **Chat bots** - Add Brummer commands to Slack/Discord ## Architecture ``` ┌─────────────────┐ JSON-RPC 2.0 ┌─────────────────┐ │ MCP Client │◄──────────────────►│ Brummer MCP │ │ (VSCode, etc) │ │ Server │ └─────────────────┘ └─────────────────┘ │ ▼ ┌─────────────────┐ │ Brummer Core │ │ - Processes │ │ - Logs │ │ - Events │ └─────────────────┘ ``` ## Key Features ### 1. Process Control Control your development processes from any MCP client: ```javascript // Start a development server await mcp.call('scripts_run', { name: 'dev' }); // Stop a process const status = await mcp.call('scripts_status', { name: 'dev' }); await mcp.call('scripts_stop', { processId: status.processId }); // Get all processes status const allStatus = await mcp.call('scripts_status'); ``` ### 2. Log Streaming Access logs in real-time: ```javascript // Search recent logs const logs = await mcp.call('logs_search', { query: 'dev', limit: 100 }); // Stream logs as they happen const stream = await mcp.call('logs_stream', { follow: true, limit: 100 }); ``` ### 3. Error Detection Get notified of errors immediately: ```javascript // Get recent errors const errors = await mcp.call('logs_search', { query: 'error', level: 'error', limit: 10 }); // Stream error events const errorStream = await mcp.call('logs_stream', { level: 'error', follow: true }); ``` ### 4. URL Management Access detected URLs: ```javascript // Get HTTP requests from proxy const requests = await mcp.call('proxy_requests', { limit: 50 }); // Open URL in browser await mcp.call('browser_open', { url: 'http://localhost:3000' }); ``` ## Quick Start ### 1. Enable MCP Server Start Brummer with MCP enabled: ```bash brum --mcp ``` Or configure in `.brum.toml`: ```toml mcp_port = 7777 no_mcp = false ``` ### 2. Connect from VSCode Add to your VSCode settings: ```json { "mcp.servers": { "brummer": { "command": "brum", "args": ["--mcp"] } } } ``` ### 3. Connect from Claude Code Add to Claude Code's MCP configuration: ```json { "mcpServers": { "brummer": { "command": "brum", "args": ["--mcp"] } } } ``` ## Common Use Cases ### 1. AI-Assisted Debugging When Claude or another AI assistant is helping you debug: ``` You: "My server is crashing, can you help?" Claude: "Let me check your server logs..." [Uses MCP to get recent errors] Claude: "I found the issue. Your server is crashing due to a missing environment variable. Here's the error: Error: Missing required environment variable: API_KEY at startServer (server.js:15:11) Would you like me to help you fix this?" ``` ### 2. Automated Testing Run tests and get results programmatically: ```javascript // Start test process const testProcess = await mcp.call('scripts_run', { name: 'test' }); // Monitor test logs const testLogs = await mcp.call('logs_stream', { processId: testProcess.processId, follow: true }); ``` ### 3. Performance Monitoring Track resource usage: ```javascript // Get telemetry sessions const sessions = await mcp.call('telemetry_sessions', { limit: 10 }); // Monitor browser performance const events = await mcp.call('telemetry_events', { eventType: 'performance', follow: true }); ``` ### 4. Build Automation Integrate with build pipelines: ```javascript // Start build const buildProcess = await mcp.call('scripts_run', { name: 'build' }); // Monitor build logs const buildLogs = await mcp.call('logs_stream', { processId: buildProcess.processId, follow: true }); ``` ## Security Considerations ### Authentication Enable authentication for production use: ```toml # Security is handled at the process level # MCP server runs locally on localhost:7777 ``` ### Access Control Limit MCP server access: ```toml # Tool access is controlled by MCP client configuration # All tools are available to connected clients ``` ### Network Security For WebSocket transport: ```toml # Brummer uses Streamable HTTP transport mcp_port = 7777 # Server only accepts local connections ``` ## Advanced Configuration ### Custom Commands Add custom MCP commands: ```json // Add custom scripts to package.json { "scripts": { "deploy": "./scripts/deploy.sh", "db:reset": "npm run db:reset" } } ``` ### Event Filtering Control which events are exposed: ```javascript // Filter events using tool parameters const errorLogs = await mcp.call('logs_stream', { level: 'error', follow: true }); const processLogs = await mcp.call('logs_stream', { processId: 'specific-process-id', follow: true }); ``` ### Rate Limiting Prevent abuse: ```toml # Rate limiting handled by HTTP server # Default limits are appropriate for development use ``` ## Troubleshooting ### MCP Server Not Starting 1. Check if another process is using the port 2. Verify Brummer has necessary permissions 3. Check logs for error messages 4. Try stdio transport instead of websocket ### Connection Issues 1. Ensure MCP server is enabled 2. Check firewall settings 3. Verify authentication tokens match 4. Enable debug logging ### Performance Issues 1. Limit log streaming rate 2. Use filters to reduce data transfer 3. Enable compression for WebSocket 4. Increase buffer sizes ## Next Steps - [API Reference](./api-reference) - Complete list of MCP methods - [Client Setup](./client-setup) - Configure various MCP clients - [Events](./events) - Real-time event documentation - [Examples](https://github.com/yourusername/brummer/tree/main/examples/mcp) - Sample implementations