Workspace MCP

A FastMCP server that exposes Google Drive, Google Sheets, and external REST API operations as tools that any MCP-compatible LLM agent (e.g. Claude Desktop) can call.

Architecture

Claude / MCP Client
        │
        │  MCP protocol
        ▼
  FastMCP Server (main.py)
        │
        ├── Google Drive API   (google_tools.py)
        ├── Google Sheets API  (google_tools.py)
        └── External REST API  (rest_tools.py)

Related MCP server: Generic MCP Server for Google Workspace

Features

• List files in a Google Drive

• Read data from a Google Sheet

• Append rows to a Google Sheet

• Search records via a external REST API

• All functionality exposed as MCP tools consumable by LLM agents

Project Structure

mcp_workspace/
│
├── main.py            # MCP server entrypoint, registers all tools
├── config.py           # Loads environment variables / settings
├── google_tools.py     # Google Drive & Sheets integration
├── rest_tools.py        # Generic external REST API integration
├── requirements.txt
├── .env.example
├── README.md
└── .gitignore

Setup

1. Clone the repo and install dependencies:

   python -m venv venv
   source venv/bin/activate   # or venv\Scripts\activate on Windows
   pip install -r requirements.txt

2. Copy .env.example to .env and fill in your own values:

   cp .env.example .env

3. Add your Google service account credentials as credentials.json in the project root

4. Run the server:

   Option 1: Using Claude Desktop (Recommended)

   Add the server to your claude_desktop_config.json:

   {
     "mcpServers": {
       "google-workspace-server": {
         "command": "python",
         "args": [
           "/path/to/main.py"
         ]
       }
     }
   }

   Replace /path/to/main.py with the absolute path to your project. Restart Claude Desktop, and the MCP server will be launched automatically.

   ---

   Option 2: Using MCP Inspector (For Development & Testing)

   Run the server with the MCP Inspector:

   npx @modelcontextprotocol/inspector python main.py

   Open the Inspector URL displayed in the terminal to view and test all available MCP tools.

Notes

• This project was generalized from an internal automation tool. The REST API integration (rest_tools.py) is written generically — swap in your own endpoint and payload shape via config.py / .env.