Best of MCP Servers — 450 Ranked MCP Servers

What it is

Best of MCP Servers is a curated, weekly-updated directory that ranks MCP (Model Context Protocol) servers. It covers 34 categories and scores each server based on GitHub stars, contributor count, and recent activity. The directory helps developers find the right MCP server for any integration need.

This resource targets developers and AI engineers building agent systems that use MCP to connect LLMs with external tools, databases, and APIs.

How it saves time or tokens

Searching for MCP servers manually means scanning GitHub, reading READMEs, and comparing quality signals yourself. This directory pre-scores and ranks servers so you can jump to the top options in your category.

The quality score combines multiple signals into a single number, saving you the effort of evaluating repository health metrics yourself.

How to use

1. Browse the directory by category (database, browser, filesystem, etc.)
2. Sort by quality score to find the most active and well-maintained servers
3. Click through to the GitHub repository for installation instructions
4. Add the server to your MCP configuration file

Example

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/dir"]
    },
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"],
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "ghp_..."
      }
    },
    "postgres": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-postgres", "postgresql://localhost/mydb"]
    }
  }
}

Related on TokRepo

• MCP Chrome integration -- Chrome browser MCP server
• MCP GitHub integration -- GitHub MCP server for repository access

Common pitfalls

• High star count does not guarantee production readiness; check the issue tracker for unresolved bugs
• Some MCP servers require specific runtime environments (Python, Node.js, Docker); verify compatibility before installing
• Quality scores are relative within categories; comparing scores across different categories is not meaningful