Model Context Protocol Servers (Archived)

This repository (modelcontextprotocol/modelcontextprotocol/servers-archived) is an archive of historical reference implementations and is no longer maintained. It does not receive security updates or bug fixes.

For current, actively maintained Model Context Protocol (MCP) servers, use the official repository:

https://github.com/modelcontextprotocol/servers

The Google Drive MCP server allows users to list, read, and search files within Google Drive. It provides access to files via specific tools and URI-based resources.

Tools

• search: Searches for files in Google Drive.
  • Input: query (string)
  • Returns: File names and MIME types of matching files.

Resources

Files can be accessed using the URI pattern gdrive:///<file_id>. The server automatically exports Google Workspace files into compatible formats:

• Docs $\rightarrow$ Markdown
• Sheets $\rightarrow$ CSV
• Presentations $\rightarrow$ Plain text
• Drawings $\rightarrow$ PNG
• Other files: Provided in their native format.

The GitHub MCP Server is an implementation of the Model Context Protocol that provides tools for interacting with the GitHub API. It enables automated file operations, repository management, issue tracking, pull request workflows, and advanced searching.

Key Features:

• Automatic Branch Creation: Automatically creates branches when creating or updating files if they do not exist.
• Git History Preservation: Performs operations that maintain proper Git history without using force pushes.
• Batch Operations: Supports pushing multiple files in a single commit via push_files.
• Advanced Search: Provides specialized tools for searching code, issues/PRs, and users using GitHub's native search syntax.

Note: Development for this project has moved to the github/github-mcp-server repository.

The server exposes two types of resources that can be read by the LLM:

1. Console Logs (console://logs): Provides browser console output in text format, including all messages from the browser.
2. Screenshots (screenshot://<name>): Provides PNG images of captured screenshots. Use the <name> provided during the puppeteer_screenshot tool call to access the specific image.

The Redis MCP server implements an exponential backoff strategy for handling connection issues:

• Max Retries: 5
• Initial Delay: 1 second
• Maximum Delay: 30 seconds

To prevent infinite reconnection loops, the server will automatically exit after reaching the maximum number of retries.

The server uses specific abstractions to facilitate data analysis and guided workflows:

Resources

• memo://insights: A dynamic resource that acts as a continuously updated business insights memo. It aggregates insights discovered during analysis and is automatically updated whenever the append_insight tool is used.

Prompts

• mcp-demo: An interactive demonstration prompt. It requires a topic argument (the business domain to analyze). It automates the setup of database schemas and sample data, then guides the user through analysis and insight generation, integrating directly with the memo://insights resource.

The server automatically discovers and provides schema information for every table in the database using the following resource URI pattern:

postgres://<host>/<table>/schema

These resources return JSON schema information, including column names and data types, derived from database metadata.

To use the Puppeteer MCP server in VS Code, add the configuration to your User Settings (JSON) or to a .vscode/mcp.json file in your workspace.

Note: If using .vscode/mcp.json, the top-level "mcp" key is not required.

NPX (Visible Browser)

Docker (Headless Chromium)

### NPX Installation (for User Settings JSON)
```json
{
  "mcp": {
    "servers": {
      "puppeteer": {
        "command": "npx",
        "args": ["-y", "@modelcontextprotocol/server-puppeteer"]
      }
    }
  }
}

Docker Installation (for User Settings JSON)

{
  "mcp": {
    "servers": {
      "puppeteer": {
        "command": "docker",
        "args": [
          "run",
          "-i",
          "--rm",
          "--init",
          "-e",
          "DOCKER_CONTAINER=true",
          "mcp/puppeteer"
        ]
      }
    }
  }
}

You can install and run the mcp-server-git using either uv (recommended) or pip.

Using uv (Recommended)

No specific installation is required. Use uvx to run the server directly.

Using PIP

Install the package via pip:

pip install mcp-server-git

After installation, you can run it as a module:

python -m mcp_server_git