Mind Map MCP Server v1.22.0

Experimental Code Intelligence Platform - A Model Context Protocol (MCP) server that explores neuroscience-inspired approaches to software development analysis. This is an experimental research project featuring advanced query caching, associative learning patterns, context awareness, attention mechanisms, temporal knowledge modeling, and code analysis tools.

⚠️ Current Status: Experimental v1.22.0

🧪 This is experimental software under active development - Use for testing and research purposes. Features may change or be removed.

Features under development: Context-aware caching • Brain-inspired learning • Code pattern detection • Document analysis • Multi-language AST parsing • File ignore patterns • CI/CD automation • Memory optimization

🎯 Latest Update v1.22.0: Cross-Language API Detection - Comprehensive API endpoint detection across 12 programming languages. New Features: Detect REST APIs (Flask, Express, Spring Boot), GraphQL schemas, gRPC services, WebSocket endpoints, and WebAssembly modules. Language Support: Python, JavaScript/TypeScript, Java, Go, Rust, C++, C#, PHP, Ruby, Swift, Kotlin, and Scala. Intelligent Detection: Framework-specific patterns with confidence scoring for accurate API discovery. New dedicated MCP tool detect_cross_language_apis for direct API analysis.

⚠️ Important Disclaimer

This is experimental software developed for research and testing purposes. It explores various approaches to code analysis and project understanding using Model Context Protocol (MCP).

Before using:

• Expect bugs, incomplete features, and breaking changes
• Use in non-production environments only
• Backup your projects before extensive use
• Report issues and provide feedback to help improve the project

Use cases:

• Improving coding workflows with Claude Code - Enhanced project understanding and context
• Experimenting with MCP server development
• Testing code analysis and pattern detection approaches
• Research into alternative programming assistance tools
• Learning about AST parsing and project structure analysis

📦 Installation & Setup

Quick Install from npm

# Install globally
npm install -g mind-map-mcp

# Or install locally in your project
npm install mind-map-mcp

Claude Code Integration

1. Automatic Setup (Recommended)

The easiest way to set up Mind Map MCP with Claude Code:

# Run the automatic setup tool
npx mind-map-mcp init-claude-code

# Or if installed globally
mind-map-mcp init-claude-code

This automatically:

• ✅ Detects your operating system and Claude installation
• ✅ Creates proper configuration files with correct paths
• ✅ Provides platform-specific setup instructions
• ✅ Includes verification commands and troubleshooting

2. Manual Setup for Claude Desktop

If you prefer manual configuration, add this to your Claude Desktop config:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%/Claude/claude_desktop_config.json Linux: ~/.config/claude-desktop/config.json

{
  "mcpServers": {
    "mind-map-mcp": {
      "command": "npx",
      "args": ["mind-map-mcp"],
      "env": {}
    }
  }
}

3. Environment Variable Configuration

You can configure the MCP server to work with specific project directories using the MCP_PROJECT_ROOT environment variable:

{
  "mcpServers": {
    "mind-map-mcp": {
      "command": "npx",
      "args": ["mind-map-mcp"],
      "env": {
        "MCP_PROJECT_ROOT": "/path/to/your/project"
      }
    }
  }
}

What it does:

• Makes MCP scan and cache files in the specified project directory
• Creates .mindmap-cache folder in the target project
• Uses project-specific configuration and mind map data
• Allows working with multiple projects independently

Usage Examples:

// For a specific project
"env": {
  "MCP_PROJECT_ROOT": "/Users/yourname/projects/my-app"
}

// For a demo or test project
"env": {
  "MCP_PROJECT_ROOT": "/Users/yourname/projects/demo-project"
}

Without this variable: MCP uses the current working directory where the MCP server was started.

4. Verify Installation

After setup, restart Claude and verify the integration:

1. Check MCP Tools: In Claude, you should see 33 new MCP tools available
2. Test Basic Functionality: Try these commands in Claude:
   Code

   Copy

   Please scan the current project and show me the statistics.
   

3. Test Features: Try experimental features:
   Code

   Copy

   Please analyze the project architecture.
   

📝 Usage with Claude Code

Once installed, you can experiment with these commands in Claude:

Sample Workflow

# Start with basic scanning:
Please scan the project and get initial statistics.

# Explore analysis features:
Please analyze the project structure and suggest areas of focus.

# Test learning features:
Please update the mind map with information about [task description].

🔬 Experimental Analysis Features

# Hebbian Learning - "Neurons that fire together, wire together"
Please show me the Hebbian learning statistics and top co-activation patterns.

# Hierarchical Context - Multi-level awareness
Please get the hierarchical context stats and most relevant context items.

# Attention System - Dynamic focus allocation
Please show the attention system statistics and allocate attention to important nodes.

# Bi-temporal Knowledge - Valid vs Transaction time tracking
Please get bi-temporal statistics and create a context window for this session.

# Pattern Prediction - Anticipatory intelligence
Please get pattern predictions and show emerging patterns.

Advanced Intelligence Features

# Get architectural insights
Please analyze the project architecture and detect design patterns.

# Find cross-language dependencies
Please detect cross-language dependencies in this polyglot project.

# Get intelligent refactoring suggestions
Please generate multi-language refactoring suggestions focused on architecture.

# Predict emerging code patterns
Please analyze and predict what code patterns are likely to emerge.

Development Tool Integration

# Detect available tools
Please detect all development tools available in this project.

# Run comprehensive analysis
Please run the full tool suite and provide aggregated results.

# Get tool recommendations
Please recommend missing development tools that would benefit this project.

📋 Verification Checklist

✅ Installation: npm list -g mind-map-mcp shows the package ✅ Claude Integration: 33 MCP tools visible in Claude ✅ Basic Functionality: scan_project command works ✅ Advanced Features: Brain-inspired tools respond correctly ✅ Multi-language Support: AST analysis works for your languages

🔧 Troubleshooting

Common Issues:

1. "No MCP tools visible" → Restart Claude after configuration
2. "Command not found" → Ensure npm global install path is in PATH
3. "Permission denied" → Run npm config get prefix and check permissions
4. "Server not responding" → Check Claude Desktop config file syntax

Get Help:

• 📖 Check QUICK_USAGE.md for workflow examples
• 📖 Read CLAUDE_CODE_SETUP.md for detailed setup
• 🐛 Report issues: https://github.com/nerfels/mind-map/issues

Features

🧠 Brain-Inspired Intelligence (Phase 6)

• Associative Memory System: Neural activation spreading across connected code concepts (50-70% relevance improvement)
• Context-Aware Query Caching: Intelligent caching with similarity matching (5-10x performance boost for repeated queries)
• Parallel Processing Engine: Chunked file analysis with worker pool orchestration (3-5x faster project scanning)
• Neuromorphic Query Patterns: Replaces linear search with brain-like associative activation networks
• Intelligent Cache Invalidation: Path-based selective cache clearing with LRU eviction and 100MB memory management
• Hebbian Learning System: Co-activation tracking with synaptic strengthening ("neurons that fire together, wire together")
• Inhibitory Learning: Failure avoidance through negative pattern recognition (30% reduction in repeated mistakes)
• Hierarchical Context Management: Multi-level context awareness (immediate, session, project, domain)
• Attention Mechanisms: Multi-modal attention fusion with cognitive load management (Miller's 7±2 rule)
• Bi-temporal Knowledge Model: Valid time vs transaction time tracking with complete audit trails
• Pattern Prediction Engine: Anticipates code patterns before they emerge using time series analysis and predictive forecasting

🚀 Memory Optimization (v1.15.0)

• Variable Lazy Loading: Intelligent memory management with 40.3% reduction in variable node memory usage
• Smart Filtering: Only loads critical variables (exported, global, heavily-used >5 references) immediately
• Summary Node System: Creates single nodes containing lazy-loaded variable metadata for thousands of variables
• On-Demand Loading: 8ms average retrieval time for pattern-based variable queries
• Full Functionality Preservation: All variable querying capabilities maintained while dramatically reducing memory footprint
• Automatic Optimization: No configuration required - automatically detects important vs. lazily-loadable variables