🚀 Claude Desktop & Windsurf IDE - Setup Guide

Enhanced MCP Memory Server with Complete ML Configuration

---

🎯 Quick Installation

Claude Desktop 🤖

# Automatic installation
curl -sSL https://raw.githubusercontent.com/PiGrieco/mcp-memory-server/feature/complete-architecture-refactor/scripts/install/install_claude.sh | bash

Windsurf IDE 🌊

# Automatic installation  
curl -sSL https://raw.githubusercontent.com/PiGrieco/mcp-memory-server/feature/complete-architecture-refactor/scripts/install/install_windsurf.sh | bash

---

🧠 What Gets Installed

✅ Complete ML Configuration:

• ML Model: PiGrieco/mcp-memory-auto-trigger-model
• Entry Point: main.py (unified modern server)
• Architecture: Enhanced server with all legacy ML features

🎯 ML Thresholds Configured:

ML_CONFIDENCE_THRESHOLD=0.7    # 70% ML confidence
TRIGGER_THRESHOLD=0.15         # 15% general trigger
SIMILARITY_THRESHOLD=0.3       # 30% search similarity  
MEMORY_THRESHOLD=0.7           # 70% memory importance
SEMANTIC_THRESHOLD=0.8         # 80% semantic similarity
ML_TRIGGER_MODE=hybrid         # Hybrid ML + rules

📚 Continuous Learning:

ML_TRAINING_ENABLED=true
ML_RETRAIN_INTERVAL=50
FEATURE_EXTRACTION_TIMEOUT=5.0
MAX_CONVERSATION_HISTORY=10
USER_BEHAVIOR_TRACKING=true
BEHAVIOR_HISTORY_LIMIT=1000

---

📋 Configuration Files Created

Claude Desktop:

• Path: ~/.config/claude/claude_desktop_config.json (Linux/Windows)
• Path: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)

Windsurf IDE:

• Path: ~/mcp-memory-server/windsurf_config.json

---

🔧 Manual Configuration

If you prefer manual setup, use these templates:

Claude Template:

{
  "mcpServers": {
    "mcp-memory-sam": {
      "command": "/path/to/mcp-memory-server/venv/bin/python",
      "args": ["/path/to/mcp-memory-server/main.py"],
      "env": {
        "ML_MODEL_TYPE": "huggingface",
        "HUGGINGFACE_MODEL_NAME": "PiGrieco/mcp-memory-auto-trigger-model",
        "AUTO_TRIGGER_ENABLED": "true",
        "CLAUDE_MODE": "true",
        "ML_CONFIDENCE_THRESHOLD": "0.7",
        "TRIGGER_THRESHOLD": "0.15",
        "SIMILARITY_THRESHOLD": "0.3",
        "MEMORY_THRESHOLD": "0.7",
        "SEMANTIC_THRESHOLD": "0.8",
        "ML_TRIGGER_MODE": "hybrid"
      }
    }
  }
}

Windsurf Template:

{
  "mcpServers": {
    "mcp-memory-sam": {
      "command": "/path/to/mcp-memory-server/venv/bin/python", 
      "args": ["/path/to/mcp-memory-server/main.py"],
      "env": {
        "ML_MODEL_TYPE": "huggingface",
        "HUGGINGFACE_MODEL_NAME": "PiGrieco/mcp-memory-auto-trigger-model",
        "AUTO_TRIGGER_ENABLED": "true",
        "WINDSURF_MODE": "true",
        "ML_CONFIDENCE_THRESHOLD": "0.7",
        "TRIGGER_THRESHOLD": "0.15",
        "SIMILARITY_THRESHOLD": "0.3",
        "MEMORY_THRESHOLD": "0.7", 
        "SEMANTIC_THRESHOLD": "0.8",
        "ML_TRIGGER_MODE": "hybrid"
      }
    }
  }
}

---

🧪 Testing Installation

After installation, verify everything works:

# Test server startup
cd ~/mcp-memory-server  
python main.py --test

# Should output: ✅ Server imports and configuration successful

---

🎉 Available SAM Tools

Once installed, you'll have access to all SAM tools:

• save_memory - Save memories with ML embeddings
• search_memory - Semantic search through memories
• analyze_message - Auto-trigger analysis with ML
• get_memory_stats - Complete system statistics
• list_memories - List all saved memories

---

🔄 Differences from Legacy

✅ Enhanced Features:

• Modern Architecture: Unified server (main.py) vs separate legacy servers
• Complete ML Config: All thresholds from legacy server included
• Better Error Handling: Improved stability and logging
• Flexible Configuration: YAML + environment variable support
• SAM Compatibility: 100% compatible with existing SAM workflows

📈 Migration Path:

• Old: cursor_mcp_server.py, claude_mcp_server.py, windsurf_mcp_server.py
• New: main.py with platform-specific environment variables
• Result: Same functionality, better architecture

---

🛠️ Troubleshooting

Common Issues:

1. Python Version: Requires Python 3.8+
2. Virtual Environment: Auto-created during installation
3. ML Model Download: ~63MB, downloaded on first use
4. MongoDB: Auto-configured for local development

Support:

For issues, check the main repository: https://github.com/PiGrieco/mcp-memory-server

---

🧠 Ready to use SAM with enhanced ML capabilities! 🚀