使用说明
🌌 Universe MCP
The world's largest, most comprehensive MCP (Model Context Protocol) library
A fully organized, searchable, and open database of all MCP servers, clients, and use cases, automatically scraped and updated from PulseMCP.com.
📊 Current Stats
- 6,488+ MCP Servers cataloged
- 3 Data categories (Servers, Clients, Use Cases)
- Daily automatic updates
- 100% open source and free
🎯 What is Universe MCP?
Universe MCP is the definitive resource for discovering and integrating Model Context Protocol (MCP) servers and tools. We automatically scrape, organize, and index all available MCP resources to create:
- ✅ The most complete MCP server directory
- ✅ Searchable and categorized data
- ✅ JSON-based data for easy integration
- ✅ Daily automatic updates
- ✅ Compatible with Claude, Perplexity, and all LLM tools
- ✅ Ready for MCP server integration
🚀 Quick Start
Clone the Repository
git clone https://github.com/willabs-ia/universe_mcp.git
cd universe_mcp
Browse Data
# View all servers
cat indexes/all-servers.json | jq '.servers[] | {name, provider, description}' | head -20
# View statistics
cat indexes/statistics.json | jq '.totals'
# Find servers by classification
cat indexes/servers-by-classification.json | jq '.classifications.official[] | .name'
Search for Specific Servers
import json
# Load all servers
with open('indexes/all-servers.json') as f:
data = json.load(f)
# Search for database servers
database_servers = [
s for s in data['servers']
if 'database' in s.get('description', '').lower()
]
print(f"Found {len(database_servers)} database-related servers")
📁 Repository Structure
universe_mcp/
├── data/ # Raw scraped data
│ ├── servers/
│ │ ├── official/ # Official MCP servers
│ │ ├── reference/ # Reference implementations
│ │ └── community/ # Community servers
│ ├── clients/ # MCP client applications
│ └── use-cases/ # Use case examples
├── indexes/ # Pre-generated search indexes
│ ├── all-servers.json # Complete server list
│ ├── all-clients.json # Complete client list
│ ├── all-usecases.json # Complete use case list
│ ├── servers-by-classification.json
│ ├── servers-by-provider.json
│ ├── servers-by-category.json
│ └── statistics.json # Statistics and metrics
├── schemas/ # JSON Schema definitions
│ ├── server.schema.json
│ ├── client.schema.json
│ └── usecase.schema.json
├── scripts/ # Automation scripts
│ ├── scrapers/ # Web scrapers + detail enrichment
│ ├── validators/ # Data validation
│ ├── indexers/ # Index generators
│ ├── search.py # CLI search tool
│ └── update.py # Main update script
├── examples/ # Usage examples
│ ├── search_servers.py # Search and filter examples
│ ├── integration_example.py # Integration with your app
│ └── README.md # Examples documentation
├── docs/ # Documentation
│ ├── INTEGRATION.md # Integration guide
│ └── API.md # API reference
├── DIARIO_PROJETO.md # Project diary (development log)
├── CONTRIBUTING.md # Contribution guidelines
└── LICENSE # MIT License
🔍 Usage Examples
Find All Official Servers
import json
with open('indexes/servers-by-classification.json') as f:
data = json.load(f)
official = data['classifications']['official']
for server in official:
print(f"{server['name']} - {server['provider']}")
Get Top Providers
import json
with open('indexes/statistics.json') as f:
stats = json.load(f)
top_providers = stats['servers']['top_providers']
for provider, count in list(top_providers.items())[:10]:
print(f"{provider}: {count} servers")
Search by Category
import json
with open('indexes/servers-by-category.json') as f:
data = json.load(f)
# List all categories
print("Available categories:", list(data['categories'].keys()))
# Get servers in a specific category
database_servers = data['categories'].get('database', [])
🛠️ Development
Prerequisites
- Python 3.11+
- pip
Installation
# Install dependencies
pip install -r requirements.txt
Run Scrapers
# Test mode (scrape only first 2 pages)
python scripts/scrapers/scrape_servers.py --test
# Full scrape
python scripts/scrapers/scrape_servers.py
# Resume from checkpoint
python scripts/scrapers/scrape_servers.py --resume
# Scrape specific page range
python scripts/scrapers/scrape_servers.py --start 10 --end 20
Validate Data
python scripts/validators/validate_data.py
Generate Indexes
python scripts/indexers/generate_indexes.py
Complete Update (All in One)
# Full update: scrape + validate + index
python scripts/update.py
# Test mode
python scripts/update.py --test
# Update only servers
python scripts/update.py --servers-only
Search Servers (CLI)
# Search by keyword
python scripts/search.py "database"
# Filter by classification
python scripts/search.py --classification official
# Filter by provider
python scripts/search.py --provider "Anthropic"
# Combine filters
python scripts/search.py "postgres" --classification community --limit 5
# Get JSON output
python scripts/search.py "ai" --json
Enrich Server Details
# Enrich servers with detailed metadata (visits individual pages)
python scripts/scrapers/enrich_server_details.py
# Test mode (enrich only 3 servers)
python scripts/scrapers/enrich_server_details.py --test
# Enrich only official servers
python scripts/scrapers/enrich_server_details.py --classification official
# Limit number of servers
python scripts/scrapers/enrich_server_details.py --limit 10
Run Examples
# Search and filter examples
python examples/search_servers.py
# Integration example
python examples/integration_example.py
🔄 Automatic Updates
This repository automatically updates daily at 3 AM UTC via GitHub Actions.
The workflow:
- Scrapes latest data from PulseMCP.com
- Validates all data against schemas
- Generates updated indexes
- Commits and pushes changes
See .github/workflows/auto-update.yml for details.
📚 Documentation
- Integration Guide - How to integrate Universe MCP with your tools
- API Reference - Complete data structure reference
- Project Diary - Development log and decisions
🔌 MCP Server Integration (Coming Soon)
Use Universe MCP as an MCP server itself:
{
"mcpServers": {
"universe-mcp": {
"command": "npx",
"args": ["-y", "@universe-mcp/server"]
}
}
}
This will allow AI agents to query the database directly with natural language.
🤝 Contributing
We welcome contributions! Here's how you can help:
Report Issues
- Found incorrect data? Open an issue
- Have suggestions? Start a discussion
Submit Data
- Fork the repository
- Add/update JSON files in
/data/ - Run validation:
python scripts/validators/validate_data.py - Submit a pull request
Improve Scripts
- Enhance scrapers for better data extraction
- Add new indexing strategies
- Improve documentation
📊 Data Quality
- ✅ All data validated against JSON Schema
- ✅ Automatic validation in CI/CD
- ✅ Daily freshness checks
- ✅ Source tracking (PulseMCP URLs)
- ✅ Timestamp tracking (
scraped_at,last_updated)
🌟 Use Cases
- LLM Tool Discovery: Help AI agents find the right MCP server for a task
- Developer Research: Explore available MCP integrations
- Ecosystem Analysis: Understand MCP adoption and trends
- Automated Integration: Build tools that auto-discover MCP capabilities
- Documentation: Reference for MCP learning resources
📜 License
MIT License - See LICENSE for details.
Feel free to use, modify, and distribute this data for any purpose.
🙏 Acknowledgments
- PulseMCP.com - Source of MCP data
- Anthropic - Creators of the Model Context Protocol
- Community - All MCP server and client developers
🔗 Links
- PulseMCP: https://www.pulsemcp.com
- MCP Specification: https://github.com/modelcontextprotocol
- Official MCP Registry: https://github.com/modelcontextprotocol/servers
📞 Contact
- Issues: https://github.com/willabs-ia/universe_mcp/issues
- Discussions: https://github.com/willabs-ia/universe_mcp/discussions
- Author: @willabs-ia
Made with ❤️ for the MCP community
