What is MCP?
The Model Context Protocol (MCP) is an open standard that enables AI applications to securely access external tools and data sources. MCP servers expose capabilities as structured tools that any MCP-compatible client can discover and invoke. Thegws mcp command starts an MCP server over stdio, exposing Google Workspace APIs as tools that AI clients (Claude Desktop, Gemini CLI, VS Code, etc.) can call directly.
Starting the MCP Server
Thegws mcp command accepts flags to control which services and tools are exposed:
Available Flags
| Flag | Description |
|---|---|
-s, --services <list> | Comma-separated list of services to expose, or all |
-w, --workflows | Also expose workflow tools |
-e, --helpers | Also expose helper tools |
Services Flag
The-s flag controls which Google Workspace services are exposed as tools. You can specify:
- Individual services:
drive,gmail,calendar,sheets,docs,slides,chat,tasks,people,admin, etc. - Multiple services:
drive,gmail,calendar - All services:
all
Workflows and Helpers
By default, only core service methods are exposed. Use-w to include workflow tools (multi-step automation sequences) and -e to include helper tools (shortcuts for common operations):
Client Configuration
MCP clients discover and connect to MCP servers via configuration files. Here’s how to configure thegws MCP server:
Claude Desktop
Add to~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%/Claude/claude_desktop_config.json (Windows):
Gemini CLI
Add to your Gemini CLI configuration:VS Code
If using an MCP extension for VS Code, add the server configuration according to the extension’s documentation.Tool Limit Considerations
Each Google Workspace service exposes roughly 10–80 tools depending on the API surface. Most MCP clients have a tool limit (typically 50–100 tools). Important: Keep your service list focused on what you actually need to stay under your client’s tool limit.Recommended Configurations
Authentication
The MCP server inherits authentication from yourgws CLI configuration. Make sure you’ve authenticated before starting the MCP server:
Next Steps
- Agent Skills - Install 100+ agent skills for AI automation
- Gemini Extension - Use the CLI with Gemini CLI
- Authentication Setup - Configure Google Cloud authentication