MCP Registry Manager 🌐

Centralized discovery and quality scoring for the MCP (Model Context Protocol) ecosystem.

Quick Start

# Install dependencies
pip install requests sentence-transformers numpy pandas
 
# Discover MCP servers
python3 scripts/mcp-registry.py --discover
 
# Semantic search
python3 scripts/mcp-registry.py --search "file system operations"
 
# Install a server
python3 scripts/mcp-registry.py --install @modelcontext/official-filesystem
 
# List installed servers
python3 scripts/mcp-registry.py --list --installed

Features

✅ Multi-source discovery — GitHub, awesome-mcp-servers, AllInOneMCP
✅ Quality scoring — Test coverage, documentation, maintenance
✅ Semantic search — Find by what it does, not just names
✅ Install management — Install/uninstall with tracking
✅ Categorization — Files, databases, APIs, dev tools, etc.

Problem

MCP is exploding (100+ servers) but:

• Discovery is fragmented
• No quality signals
• No semantic search
• No unified management

Quality Score

Categories

• Files (filesystem, S3)
• Databases (PostgreSQL, MongoDB, Redis)
• APIs (HTTP, GraphQL, REST)
• Dev Tools (Git, Docker, CI)
• Media (image, video, audio)
• Utilities (time, crypto, encryption)

Installation

git clone https://github.com/orosha-ai/mcp-registry-manager
pip install requests sentence-transformers numpy pandas

License

MIT