Durandal - Universal AI Memory System

Durandal is an advanced AI memory system providing localized, on-premise persistent memory for any AI assistant (Claude, ChatGPT, Gemini). Features seamless Claude Code integration via MCP (Model Context Protocol), intelligent caching with RAMR, and sophisticated context management.

🚀 Quick Start

Installation

npm install -g @enterprise/durandal-memory

Basic Setup

1. Initialize Configuration

   durandal --init

2. Set Up Database (Automated)

   durandal --setup-db

3. Configure API Key

   # Edit .env file with your Claude API key
   CLAUDE_API_KEY=your_api_key_here

4. Start Durandal

   durandal

📋 Features

Core Capabilities

• 🧠 Persistent Memory: Maintains conversation context across sessions
• 🔍 Semantic Search: AI-powered content discovery and retrieval
• ⚡ RAMR Cache: Rapid Access Memory Register for instant responses
• 🌐 Multi-AI Support: Works with Claude, OpenAI, Google AI, and more
• 🔧 MCP Integration: Native Claude Code support via Model Context Protocol
• 📊 Knowledge Analytics: Intelligent pattern recognition and insights

Advanced Features

• Context Management: Intelligent conversation context optimization
• Knowledge Extraction: Automatic insight generation from conversations
• Cross-Project Memory: Shared knowledge across multiple projects
• Real-time Optimization: Automatic memory system tuning
• Enterprise Ready: PostgreSQL support, backup/recovery, monitoring

🛠️ Installation Options

1. NPM Global Installation (Recommended)

npm install -g @enterprise/durandal-memory

# Available commands after installation:
durandal          # Main CLI interface
durandal-ui       # Web UI server
durandal-mcp      # MCP server for Claude Code

2. Local Development Installation

git clone <repository-url>
cd durandal-memory
npm install
npm start

3. Claude Code MCP Integration

# Install globally first
npm install -g @enterprise/durandal-memory

# Add to Claude Code MCP configuration
durandal-mcp --configure-claude-code

⚙️ Configuration

Environment Setup

Create .env file from template:

cp .env.distribution .env

Required Configuration

# AI Provider (Choose one or more)
CLAUDE_API_KEY=your_claude_api_key_here
OPENAI_API_KEY=your_openai_api_key_here
GOOGLE_API_KEY=your_google_api_key_here

# Database (SQLite default, PostgreSQL for enterprise)
DATABASE_TYPE=sqlite
DATABASE_PATH=./durandal-memory.db

# Basic Settings
NODE_ENV=production
PORT=3000

Database Options

Automated Setup (Recommended)

durandal --setup-db

Interactive wizard handles everything automatically.

SQLite (Default - Zero Configuration)

DATABASE_TYPE=sqlite
DATABASE_PATH=./durandal-memory.db

PostgreSQL (Enterprise)

DATABASE_TYPE=postgresql
DATABASE_URL=postgresql://username:password@localhost:5432/durandal_memory

🎯 Usage

CLI Mode

durandal
# Interactive CLI with full AI conversation and memory features

UI Mode

durandal-ui
# Web interface at http://localhost:3000

MCP Server Mode (Claude Code Integration)

durandal-mcp
# Starts MCP server for Claude Code integration

Programmatic Usage

const { DurandalAI } = require('@enterprise/durandal-memory');

const durandal = new DurandalAI({
    provider: 'claude',
    apiKey: process.env.CLAUDE_API_KEY,
    databaseType: 'sqlite'
});

await durandal.initialize();
const response = await durandal.chat("Remember my preference for React with TypeScript");

🔧 Claude Code Integration

Automatic Setup

durandal-mcp --configure-claude-code

Manual Setup

1. Add to Claude Code MCP Configuration (~/.claude/settings.json):

{
  "mcpServers": {
    "durandal-memory": {
      "command": "durandal-mcp",
      "args": [],
      "env": {}
    }
  }
}

2. Available MCP Tools:
   • durandal-memory__store_memory
   • durandal-memory__search_memories
   • durandal-memory__get_context
   • durandal-memory__optimize_memory

📚 Commands & Features

Memory Commands

• store <content> - Store information in memory
• search <query> - Search stored memories
• recall <topic> - Recall specific topics
• optimize - Optimize memory system

Context Commands

• context - Show current context
• clear - Clear current session
• sessions - List all sessions
• export - Export memory data

System Commands

• status - System health check
• stats - Memory statistics
• backup - Create backup
• restore - Restore from backup

🏗️ Architecture

Core Components

• DurandalAI: Main orchestrator and CLI interface
• ClaudeClient: AI provider abstraction layer
• RAMR: Rapid Access Memory Register (intelligent caching)
• ContextManager: Conversation context optimization
• KnowledgeAnalyzer: Pattern recognition and insights
• DatabaseAdapter: Multi-database support layer

Memory Layers

1. Hot Cache: In-memory for instant access
2. RAMR Cache: SQLite-based persistent cache
3. Main Database: Full conversation history
4. Knowledge Base: Extracted insights and patterns

🔒 Security & Privacy

• Local Processing: All data remains on your machine
• Encrypted Storage: Database encryption support
• API Key Security: Secure credential management
• Access Control: User authentication and authorization
• Audit Logging: Complete activity tracking

📊 Performance

• Sub-second Response: RAMR cache for instant memory access
• Scalable Architecture: Handles millions of memories
• Optimized Queries: Intelligent database indexing
• Memory Efficiency: Automatic cleanup and optimization
• Resource Monitoring: Built-in performance metrics

🧪 Testing

# Run all tests
npm run test-all

# Specific test suites
npm run test-db        # Database integration
npm run test-memory    # Memory system
npm run test-context   # Context management
npm run test-ramr      # RAMR cache system

🐛 Troubleshooting

Common Issues

Database Connection

durandal --test-setup
# Validates database connectivity

durandal --setup-db
# Re-run database setup wizard

API Key Issues

durandal --validate-keys
# Checks all configured API keys

Memory Performance

durandal --optimize
# Runs memory optimization

Debug Mode

DEBUG=durandal:* durandal
# Enables verbose logging

🤝 Contributing

1. Fork the repository
2. Create feature branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🆘 Support

• Documentation: Full Documentation
• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: support@durandal-ai.com

🗺️ Roadmap

Version 2.2.0 (Coming Soon)

• Multi-user support
• Team collaboration features
• Advanced analytics dashboard
• Mobile application

Version 2.3.0 (Future)

• Blockchain memory verification
• Advanced AI model training
• Enterprise SSO integration
• Cloud synchronization options

Made with ❤️ by Enterprise Data and Analytics

Durandal - Empowering AI with Perfect Memory