Enhanced Dash MCP Server
An intelligent Model Context Protocol (MCP) server that transforms your local Dash documentation into a powerful, context-aware assistant for Claude. Built for developers who want seamless access to documentation while coding.
🚀 Features
Core Capabilities
- 🔍 Intelligent Search - Fuzzy matching with typo tolerance and smart ranking
- 📚 Content Extraction - Clean text extraction from HTML, Markdown, and text docs
- ⚡ Multi-Tier Caching - Memory + disk caching for lightning-fast repeated searches
- 🎯 Project Awareness - Automatically detects your tech stack and prioritizes relevant docs
- 🛠️ Implementation Guidance - Best practices and patterns for specific features
- 📈 Migration Support - Version upgrade documentation and breaking changes
- 🔄 Latest API Reference - Current API docs with practical examples
Developer Workflow Integration
- Warp Terminal - Native command palette and workflow integration
- tmux - Background server execution across terminal sessions
- Neovim - Documentation access while coding via Claude
- Oh-My-Zsh - Enhanced aliases and productivity shortcuts
- Git Integration - Repository-aware documentation suggestions
Supported Technologies
JavaScript/TypeScript, React, Next.js, Vue.js, Angular, Node.js, Python, Django, Flask, FastAPI, pandas, NumPy, and many more through Dash docsets.
📋 Prerequisites
- macOS with Dash app installed
- Python 3.8+
- Dash docsets downloaded (JavaScript, Python, React, etc.)
- Claude with MCP support
- tmux (recommended for background execution)
⚡ Quick Start
1. Clone & Setup
# Clone or download the project files
mkdir ~/enhanced-dash-mcp && cd ~/enhanced-dash-mcp
# Make setup script executable
chmod +x setup-dash-mcp.sh
# Run automated setup
./setup-dash-mcp.sh
Add this to Claude's MCP settings:
{
"mcpServers": {
"enhanced-dash-mcp": {
"command": "python3",
"args": [
"/Users/your-username/mcp-servers/enhanced-dash-mcp/enhanced_dash_server.py"
],
"env": {}
}
}
}
3. Start & Test
# Add shell enhancements
echo "source ~/mcp-servers/enhanced-dash-mcp/dash-mcp-aliases.sh" >> ~/.zshrc
source ~/.zshrc
# Start the server
dash-mcp-start
# Test with Claude
# "Search for React useState hook documentation"
🎮 Usage
Basic Documentation Search
# Ask Claude:
"Search for Python pandas DataFrame methods"
"Find React hooks best practices"
"Get FastAPI routing documentation with examples"
Project-Aware Intelligence
# Navigate to your project directory, then ask Claude:
"Analyze my current project and find relevant documentation"
"Get implementation guidance for user authentication in my React app"
"What are the best practices for my current Django project?"
Migration & Upgrade Help
# Ask Claude:
"Get migration docs for upgrading from React 17 to 18"
"Find Django 4.2 upgrade guide and breaking changes"
"Show me Next.js 13 to 14 migration documentation"
API Reference with Examples
# Ask Claude:
"Get latest pandas DataFrame.merge API reference with examples"
"Show me React useEffect hook documentation and patterns"
"Find Express.js middleware documentation with use cases"
🛠️ Advanced Setup
Warp Terminal Integration
For enhanced Warp Terminal support:
# Run Warp-specific setup
chmod +x setup-warp-dash-mcp.sh
./setup-warp-dash-mcp.sh
# Use Command Palette (⌘K):
dash-mcp-start
dash-analyze-project
dash-api-ref useState react
Shell Aliases & Functions
After setup, you'll have these convenient commands:
dash-mcp-start # Start server in tmux
dash-mcp-status # Check if running
dash-mcp-logs # View server output
enhanced-dash-mcp-for-project # Analyze current project
dash-api-lookup <api> <tech> # Quick API reference
dash-best-practices <feature> # Implementation guidance
dash-help # Show all commands
Powerlevel10k Integration
Add MCP server status to your prompt:
# Add to ~/.p10k.zsh (see p10k-dash-mcp.zsh for details)
# Shows 📚 when running, 📕 when stopped
🔧 Configuration
Cache Settings
# Default cache TTL: 1 hour
# Cache location: ~/.cache/dash-mcp/
# Memory + disk caching for optimal performance
Fuzzy Search Tuning
# Default threshold: 60% match
# Adjustable in server configuration
# Typo tolerance with intelligent ranking
# Default: 5000 characters per document
# Configurable for performance vs. detail trade-off
🏗️ Architecture
Core Components
- DashMCPServer - Main server orchestrating all components
- CacheManager - Multi-tier caching (memory + disk)
- ContentExtractor - Clean text extraction from various formats
- FuzzySearchEngine - Intelligent search with ranking algorithms
- ProjectAwareDocumentationServer - Context-aware documentation selection
Data Flow
- Query received from Claude via MCP
- Project context analyzed (language, framework, dependencies)
- Relevant docsets identified and prioritized
- Fuzzy search performed with intelligent ranking
- Content extracted and cached for future requests
- Results returned with project-specific scoring
Caching Strategy
- Memory Cache - Instant access for recently searched items
- Disk Cache - Persistent storage surviving server restarts
- Smart Expiration - 1-hour TTL with automatic cleanup
- Cache Keys - Generated from search parameters for optimal hit rates
Benchmarks
- First search: ~500ms (includes docset scanning)
- Cached searches: ~50ms (memory cache hits)
- Content extraction: +200-300ms (when requested)
- Fuzzy matching: Minimal overhead with significant quality improvement
Optimization Tips
- Keep server running in tmux for best performance
- Initial searches per docset are slower (cache building)
- Content extraction adds latency but provides much richer context
- Memory cache provides fastest repeated access
| Tool | Description | Use Case |
|---|
search_dash_docs | Basic documentation search with fuzzy matching | General API/concept lookup |
list_docsets | Show all available documentation sets | Discover available documentation |
get_doc_content | Get full content for specific documentation | Deep dive into specific topics |
| Tool | Description | Use Case |
|---|
analyze_project_context | Detect project tech stack and dependencies | Understand current project |
get_project_relevant_docs | Context-aware documentation search | Find docs relevant to your project |
get_implementation_guidance | Best practices for specific features | Implementation planning |
| Tool | Description | Use Case |
|---|
get_migration_docs | Version upgrade documentation | Planning upgrades and migrations |
get_latest_api_reference | Current API docs with examples | Quick reference while coding |
🚨 Troubleshooting
Common Issues
❌ "No docsets found"
# Ensure Dash is installed with docsets
ls ~/Library/Application\ Support/Dash/DocSets/
# Should show *.docset directories
❌ "Permission errors"
# Check Python environment
which python3
source ~/mcp-servers/enhanced-dash-mcp/venv/bin/activate
❌ "Import errors"
# Reinstall dependencies
cd ~/mcp-servers/enhanced-dash-mcp
source venv/bin/activate
pip install -r requirements.txt
❌ "Server won't start"
# Check if port is in use
tmux kill-session -t dash-mcp
dash-mcp-start
❌ "Slow searches"
# First searches build cache - subsequent searches are much faster
# Check cache directory
ls ~/.cache/dash-mcp/
Debug Mode
# View detailed server logs
dash-mcp-logs
# Attach to server session for real-time debugging
dash-mcp-attach
🤝 Contributing
Development Setup
# Clone repository
git clone <repository-url>
cd enhanced-dash-mcp
# Create development environment
python3 -m venv dev-env
source dev-env/bin/activate
pip install -r requirements.txt
# Install development dependencies
pip install pytest black flake8 mypy
Running Tests
# Unit tests
pytest tests/
# Linting
black .
flake8 .
mypy .
Adding New Features
- Docset Support - Add new file format extractors in
ContentExtractor - Search Algorithms - Enhance ranking in
FuzzySearchEngine - Project Detection - Extend framework detection in
ProjectAwareDocumentationServer - Caching Strategies - Optimize cache management in
CacheManager
📄 License
MIT License - see LICENSE file for details.
🙏 Acknowledgments
- Dash by Kapeli for providing excellent local documentation
- Anthropic for Claude and the MCP framework
- Warp Terminal for innovative terminal experience
- Fort Collins Tech Community for inspiration and feedback
📞 Support
- Issues: Open a GitHub issue for bugs or feature requests
- Discussions: Use GitHub Discussions for questions and ideas
- Documentation: Check the
/docs directory for detailed guides
🗺️ Roadmap
v1.1 - Enhanced Intelligence
v1.3 - Advanced Features
Built with ❤️ in Fort Collins, CO for developers who value efficient, intelligent documentation access.
Transform your development workflow with context-aware documentation that understands your project and coding patterns.