ACE MCP Server

# 🎯 Setting Up ACE MCP in Cursor AI ## 📋 Step-by-Step Instructions ### 1. 🚀 Starting ACE MCP Server First, make sure the ACE MCP server is running locally: ```bash cd $HOME/code/perplexity/ace-mcp-server docker-compose -f docker-compose.dev.yml up -d ``` Check status: ```bash curl http://localhost:34301/health ``` ### 2. 🔧 Cursor AI Configuration #### Option A: Through Cursor Settings 1. Open **Cursor AI** 2. Go to **Settings** (⌘ + ,) 3. Find **MCP Servers** or **Extensions** section 4. Add a new MCP server with settings: ```json { "name": "ACE Context Engineering", "command": "npx", "args": ["tsx", "src/index.ts"], "cwd": "$HOME/code/perplexity/ace-mcp-server", "env": { "LLM_PROVIDER": "deepseek", "API_BEARER_TOKEN": "token", "ACE_SERVER_PORT": "34301", "ACE_CONTEXT_DIR": "$HOME/code/perplexity/ace-mcp-server/contexts", "ACE_MAX_PLAYBOOK_SIZE": "10000" } } ``` #### Option B: Through Configuration File 1. Find Cursor configuration file: - **macOS**: `~/Library/Application Support/Cursor/User/settings.json` - **Linux**: `~/.config/Cursor/User/settings.json` - **Windows**: `%APPDATA%\Cursor\User\settings.json` 2. Add MCP settings to the file: ```json { "mcp.servers": { "ace-context-engineering": { "command": "npx", "args": ["tsx", "src/index.ts"], "cwd": "$HOME/code/perplexity/ace-mcp-server", "env": { "LLM_PROVIDER": "deepseek", "API_BEARER_TOKEN": "token", "ACE_SERVER_PORT": "34301", "ACE_CONTEXT_DIR": "$HOME/code/perplexity/ace-mcp-server/contexts", "ACE_MAX_PLAYBOOK_SIZE": "10000" } } } } ``` ### 3. 🔄 Restart Cursor AI After adding configuration: 1. Fully close Cursor AI 2. Start it again 3. Check logs to verify MCP server connected ### 4. ✅ Connection Verification In Cursor AI, try using commands: ``` @ace-context-engineering generate a Python function for sorting ``` or ``` @ace-context-engineering reflect on this code: def sort_list(arr): return sorted(arr) ``` ## 🎯 Available ACE MCP Commands ### 📝 Generator ``` @ace generate <prompt> ``` Generates development trajectories based on prompt. ### 🧠 Reflector ``` @ace reflect <code> ``` Analyzes code and creates insights for improvement. ### 📚 Curator ``` @ace curate <insights> ``` Updates Playbook with new knowledge. ### 📖 Playbook ``` @ace playbook ``` Shows current Playbook with patterns and best practices. ## 🔧 Environment Settings Make sure dependencies are installed: ```bash # In project directory npm install -g tsx npm install ``` ## 🐛 Troubleshooting ### Problem: MCP Server Not Connecting **Solution**: 1. Check that Docker containers are running 2. Make sure port 34301 is free 3. Verify path correctness in configuration ### Problem: Authentication Error **Solution**: 1. Check Bearer token in `.env` file 2. Ensure token matches in Cursor configuration ### Problem: Commands Not Working **Solution**: 1. Restart Cursor AI 2. Check MCP server logs: `docker logs ace-mcp-server-dev` 3. Make sure you're using correct command syntax ## 📊 Monitoring To monitor ACE MCP operation: 1. **Dashboard**: http://localhost:34300 2. **API Status**: http://localhost:34301/health 3. **Docker Logs**: `docker logs -f ace-mcp-server-dev` ## 🚀 Ready! After setup, ACE MCP will: - ✅ Generate development trajectories - ✅ Analyze your code - ✅ Accumulate knowledge in Playbook - ✅ Suggest improvements based on experience Now Cursor AI will use self-improving context for better development assistance! 🎉