Where to Find the Directory of MCP Servers: A Complete Guide

The authoritative directory of Model Context Protocol (MCP) servers is maintained in the punkpeye/awesome-mcp-servers repository and accessible via the web interface at https://glama.ai/mcp/servers or directly in the repository's README.md file.

The awesome-mcp-servers repository serves as the community-curated index for MCP server implementations. Whether you are looking for filesystem integrations, database connectors, or AI service wrappers, this directory provides the definitive list of available servers compatible with the Model Context Protocol.

Accessing the Directory of MCP Servers

There are two primary methods to explore the available MCP servers: through the synchronized web interface or by examining the source files directly.

Using the Web-Based Directory at Glama.ai

The fastest way to browse the directory of MCP servers is through the automated web interface hosted at https://glama.ai/mcp/servers. This page is automatically generated from the repository's source files, ensuring the listings remain current with every update to the main branch.

According to the source code in README.md at line 76, the repository explicitly references this web directory with the statement: "We now have a web-based directory … that is synced with the repository." Any server added to the Markdown list under the Server Implementations section in the README appears immediately on the Glama.ai interface.

Browsing the Repository Directly

For programmatic access or offline viewing, navigate to the README.md file in the root of the punkpeye/awesome-mcp-servers repository. This file contains the master list of all MCP server implementations organized by category, including direct links to each server's source code and documentation.

How to Retrieve MCP Server Entries

You can access the directory data using command-line tools or standard web browsing. Both methods return the same canonical dataset, though the CLI approach allows for local filtering.

Web Browser Access

Open the synchronized directory directly in your default browser using standard shell commands:

# Open the searchable web directory

open https://glama.ai/mcp/servers

Command-Line Interface (CLI) Method

The repository ecosystem includes a helper CLI for local searching. Install the tool globally via npm to query the directory without a browser:

# Install the CLI utility

npm install -g awesome-mcp-tools

# Search for specific server types (e.g., Python-based implementations)

awesome-mcp search python

The CLI parses the local Markdown list from README.md and filters entries based on your keyword criteria, returning results identical to the web interface.

Repository Structure and Key Files

Understanding the file structure helps when contributing new servers or accessing localized versions. The following files define and maintain the MCP servers directory:

• README.md – The primary documentation file containing the Server Implementations section and the link to the web directory at line 76.

• README-zh.md, README-zh_TW.md, README-ja.md – Translations of the main directory into Chinese (Simplified), Chinese (Traditional), and Japanese, respectively, for non-English readers.

• CONTRIBUTING.md – Guidelines and requirements for submitting new MCP servers to the curated list.

• LICENSE – MIT license governing the repository content.

These files are located in the repository root and together constitute the complete definition of the directory.

Summary

• The punkpeye/awesome-mcp-servers repository maintains the definitive directory of MCP servers.
• https://glama.ai/mcp/servers provides a searchable, auto-synced web view of the directory.
• The canonical list resides in README.md, referenced at line 76.
• Use awesome-mcp-tools CLI for local, programmatic searches of the directory.
• Translated versions are available in README-zh.md, README-zh_TW.md, and README-ja.md.

Frequently Asked Questions

Is the Glama.ai directory official or community-maintained?

The Glama.ai directory is automatically generated from the official awesome-mcp-servers repository. As stated in the README.md source, the web interface stays synchronized with the repository, making it an official mirror of the community-curated list.

How do I add my MCP server to the directory?

Submit a pull request to the punkpeye/awesome-mcp-servers repository following the guidelines in CONTRIBUTING.md. Once merged into the main branch, your server appears in both the README.md and the Glama.ai web directory within minutes.

Can I access the directory without using the web interface?

Yes. Clone the repository locally and read the README.md file directly, or install the awesome-mcp-tools CLI via npm to search and filter servers from your terminal without opening a browser.

Are the translations of the directory kept up to date?

The Chinese (Simplified, Traditional) and Japanese translations in files like README-zh.md and README-ja.md are maintained by community contributors. While they aim to sync with the main README.md, the English version in the root README remains the authoritative source for the most current server listings.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}