Curated MCP Server Directory by Stephen Akinyemi

Overview

The Curated MCP Server Directory by Stephen Akinyemi is a community-maintained collection of Model Context Protocol (MCP) servers. It gathers server endpoints, descriptions, ratings, and community notes to help developers discover reliable MCP hosts, compare features, and join active communities. The repository centralizes listings and includes contribution guidance so that server owners and users can keep entries up to date.

For developers building integrations, experiments, or client tooling that rely on MCP endpoints, this directory reduces discovery friction. You can quickly find servers that match capability, latency, cost, or community size, and you can programmatically consume the listing for automated tooling, health checks, or client configuration.

Features

• Curated list of MCP servers with metadata (address, tags, description, rating)
• Community reviews and notes for real-world reliability feedback
• Structured listing format (JSON/YAML) suitable for programmatic use
• Contribution workflow using GitHub issues and pull requests
• Example client and scripts for validating server availability
• Option to run locally for browsing and testing or to integrate into CI/CD

Installation / Configuration

Clone the repository and run the local site or tooling. The exact commands depend on the tooling in the repo (static site, Node app, or scripts). Examples below assume a Node-based static site or CLI utilities.

Clone repo:

git clone https://github.com/appcypher/awesome-mcp-servers.git
cd awesome-mcp-servers

Install dependencies (Node.js):

npm install

Run development server:

npm run dev
# or
npm start

Build for production:

npm run build

Docker (if provided):

docker build -t mcp-directory .
docker run -p 8080:8080 mcp-directory
# or with docker-compose
docker compose up --build

Environment variables (.env.example):

NODE_ENV=development
PORT=8080
LISTING_FILE=./data/servers.json
# Optional: API_KEY for protected operations
API_KEY=your_api_key_here

Repository layout (example):

Available Resources

• listings: A machine-readable file (JSON/YAML) containing server entries. Typical fields:

{
  "name": "Example MCP Server",
  "address": "https://mcp.example.org",
  "tags": ["public", "low-latency", "chat"],
  "rating": 4.7,
  "contact": "[email protected]",
  "notes": "Good for prototyping, rate-limited for free users"
}

• health-check scripts: Small utilities to ping endpoints and report status.
• example clients: Minimal examples showing how to connect to an MCP server from a client library.
• contribution guidelines: Steps for adding or updating a server, including required metadata and verification tips.

Use Cases

1. Discover an MCP server for development

   • Problem: You need a public MCP endpoint with low latency for testing.
   • Steps: Browse listings, filter by “low-latency” tag, check ratings and recent notes, then use the example client to connect.

2. Automate health checks in CI

   • Problem: Ensure production MCP endpoints are reachable before deploying.
   • Steps: Add a health-check script from scripts/ to your CI pipeline; point it to the JSON listing for target addresses and fail builds on timeout.

3. Build a client that selects a server by capability

   • Problem: Choose a server that supports a specific model or feature.
   • Steps: Fetch the listing JSON programmatically, filter by desired tag or capability field, and configure the client to use the selected address.

4. Join and evaluate communities

   • Problem: Assess community activity before joining a server.
   • Steps: Read “notes” and “contact” fields in the listing; follow the contact link or community URL to join chats and ask questions.

Contributing

• Fork the repo, add or update an entry in data/servers.json (or the repo’s canonical listing), and submit a pull request.
• Include:
  • accurate address (URL)
  • short description and tags
  • contact or source link
  • evidence of server reliability (uptime notes, benchmark results)
• Run any provided validation script:

npm run validate
# or
python scripts/validate_listings.py

• Open an issue if you need to discuss a server addition or removal.

Notes and Best Practices

• Treat public MCP servers as untrusted execution environments—avoid sending sensitive data.
• Verify rate limits and usage policies before integrating a server into production.
• Use programmatic listings and health checks to select the most appropriate endpoint for your workload.

For the latest listings and contribution instructions, see the GitHub repository: https://github.com/appcypher/awesome-mcp-servers