sow-mcp-server

williamjvest/sow-mcp-server

3.2

If you are the rightful owner of sow-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

The Model Context Protocol (MCP) server is designed for intelligent search and retrieval of over 1,390 Statement of Work (SOW) documents, offering ultra-fast search capabilities and AI context generation.

Tools
7
Resources
0
Prompts
0

SOW MCP Server

Model Context Protocol (MCP) server for intelligent search and retrieval of 1,390+ Statement of Work documents.

Status: ✅ COMPLETE AND PRODUCTION-READY

Fully functional MCP server with comprehensive testing completed. Provides sub-100ms search times across entire SOW database.

Features

  • Ultra-Fast Search: ripgrep-powered full-text search (59ms average)
  • Metadata Filtering: Filter by client, project type, budget, dates
  • AI Context Generation: Claude-ready formatted context
  • Comprehensive Database: 1,390 SOWs from 454 clients
  • Async Architecture: Full async/await support
  • Type Safety: Complete type hints and validation

Quick Start

# Install dependencies
pip install mcp
brew install ripgrep

# Start server
python3 start_server.py

# Test functionality
python3 integration_test.py

Architecture

  • Database Path: ~/Projects/sow-database/ (external reference)
  • Search Engine: ripgrep for full-text search
  • Metadata Cache: In-memory JSON index
  • MCP Tools: 7 comprehensive tools for search/retrieval

MCP Tools Available

  1. search_sows - Full-text search with filters
  2. filter_sows - Metadata-only filtering
  3. get_sow - Retrieve specific SOW by ID
  4. list_clients - List all unique clients
  5. list_project_types - Get available project types
  6. get_database_stats - Database statistics
  7. get_claude_context - AI-ready context generation

Performance Metrics

  • Database Size: 1,390 documents (27MB)
  • Search Speed: 59ms average query time
  • Memory Usage: ~50MB for full cache
  • Startup Time: <1 second
  • Test Coverage: 100% pass rate

Dependencies

  • MCP Library: Model Context Protocol server framework
  • ripgrep: Ultra-fast text search engine
  • External Database: References ~/Projects/sow-database/

Integration Notes

This server references the SOW database externally at ~/Projects/sow-database/. The database contains:

  • 1,390 markdown SOW files with YAML frontmatter
  • JSON metadata index (1.2MB)
  • Organized by client in 445 folders

The database is backed up separately in the sow-database GitHub repository.

Testing

Comprehensive test suite included:

  • integration_test.py - Full functionality test
  • test_mcp_client.py - MCP client testing
  • test_search.py - Search engine validation

All tests pass with 100% success rate.

Deployment

Ready for production deployment with Claude Desktop or any MCP-compatible client.