🚀 Google VPS MCP
Google VPS MCP is a professional Model Context Protocol (MCP) server that enables AI assistants to manage your Google Cloud VPS instances securely and efficiently. Built with TypeScript and designed for production use, it provides seamless integration with ChatGPT, Claude Desktop, and any MCP-compatible AI tool.
Complete VPS Setup
Automated VPS configuration with security hardening, Node.js, PM2, and firewall rules.
One-Click Deploy
Deploy directly from GitHub with automatic dependency installation and PM2 configuration.
Live Logs
Stream logs in real-time, search through history, and get AI-powered troubleshooting.
Secure Environment
Manage .env files securely with validation, backup, and masked sensitive values.
⚡ Quick Start (3 Steps)
1. Install & Build
Clone the repository and build:
git clone https://github.com/dannymaaz/google-vps-mcp.git
cd google-vps-mcp
npm install
npm run build⚠️ This is example code. Follow the installation steps above.
2. Authenticate with Google Cloud
Login to your Google Cloud account:
gcloud auth application-default loginThis opens your browser. Sign in with your Google account.
3. Configure Your AI Tool
Add ONE configuration to your AI tool (ChatGPT, Claude, etc.):
{
"mcpServers": {
"google-vps": {
"command": "node",
"args": ["/YOUR_PATH/google-vps-mcp/dist/index.js"],
"env": {}
}
}
}⚠️ Replace /YOUR_PATH/ with your actual path. No project ID needed!
✨ One Configuration for All AI Tools
This single JSON configuration works with ChatGPT, Claude Desktop, and any MCP-compatible AI tool. You don't need separate configurations for each tool.
📦 Installation
Prerequisites
Make sure you have:
- Node.js 18.0.0 or higher
- Google Cloud SDK (gcloud)
- Git
Clone & Install
git clone https://github.com/dannymaaz/google-vps-mcp.git
cd google-vps-mcp
npm install
npm run build⚙️ Configuration
Google Cloud Authentication
Authenticate with Google Cloud:
gcloud auth application-default login📝 Important Notes
- No Project ID Required: The MCP lists your projects automatically
- Dynamic Selection: Choose projects when creating resources
- Multiple Projects: Work with all your Google Cloud projects
Environment Configuration (Optional)
Create a .env file if you want to set defaults:
# Default Zone (optional)
GOOGLE_CLOUD_ZONE=us-central1-a
# GitHub Credentials (optional, for private repos)
GITHUB_USERNAME=your-username
GITHUB_APP_PASSWORD=your-app-password⚠️ This is an example. Copy to .env and fill in your values.
🖥️ VPS Setup
Initialize and configure a new Google Cloud VPS from scratch. Sets up SSH, firewall rules, and basic security.
Example Conversation with AI:
Parameters
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
projectId |
string | ✅ | - | Google Cloud project ID or name |
instanceName |
string | ✅ | - | Name for the VPS instance |
zone |
string | ❌ | us-central1-a | GCP zone |
machineType |
string | ❌ | e2-medium | Machine type |
diskSizeGb |
number | ❌ | 20 | Disk size in GB |
🚀 Deploy
Deploy a project from GitHub to the VPS. Clones repo, installs dependencies, and starts the application.
Example Conversation:
📋 Logs
View Logs Live
Stream live logs from the VPS in real-time. Perfect for monitoring deployments and debugging.
Example:
🔐 Environment Management
Configure Environment
Create or update .env file on the VPS. Securely manage environment variables for production.
🔒 Security Features
- Sensitive values are automatically masked
- Backup created before changes
- File permissions set to 600 (owner only)
- Validation before applying changes
🤖 AI Integration
Google VPS MCP works with any MCP-compatible AI tool. Use ONE configuration for all:
Universal Configuration
{
"mcpServers": {
"google-vps": {
"command": "node",
"args": ["/absolute/path/to/google-vps-mcp/dist/index.js"],
"env": {}
}
}
}⚠️ Replace the path with your actual path to dist/index.js
Supported AI Tools
ChatGPT
With MCP-compatible client like ChatGPT Desktop
Claude Desktop
Native MCP support built-in
Google Antigravity
Google's IDE with MCP support
OpenCode
MCP-compatible code assistant
Any MCP Client
If it supports MCP protocol, it works!
How to Setup
Copy the Configuration
Copy the universal configuration JSON above
Open Your AI Tool Settings
Go to settings or preferences in your AI tool
Paste and Update Path
Paste the JSON and replace the path with your actual path
Restart Your AI Tool
Close and reopen your AI tool to load the new configuration
🏗️ Architecture
Google VPS MCP follows a client-server architecture based on the Model Context Protocol (MCP), designed to be secure, scalable, and cross-platform.
Architecture Diagram
Figure 1: Complete system architecture showing all layers and components
Key Components
Client Layer
User interfaces including ChatGPT, Claude Desktop, Google Antigravity, and other MCP-compatible clients.
MCP Server Layer
Core server handling tool requests, security validation, project resolution, and GitHub credentials.
Service Layer
Business logic including VPS Manager, Deploy Manager, Log Manager, and Env Manager.
Infrastructure Layer
External services: Google Cloud Platform (Compute Engine, Logging, IAM) and GitHub.
❓ FAQ
What AI tools support this MCP?
Google VPS MCP works with any AI tool that supports the Model Context Protocol (MCP), including ChatGPT, Claude Desktop, Google Antigravity, OpenCode, and other MCP-compatible clients.
Do I need to configure a project ID?
No! The MCP automatically lists your Google Cloud projects. You choose which project to use when creating resources. This allows you to work with multiple projects without reconfiguration.
Is my data secure?
Yes! The MCP includes multiple security layers: command validation, environment variable masking, secure SSH connections, and Google Cloud's enterprise-grade security.
Can I use this with multiple VPS instances?
Absolutely! You can manage multiple VPS instances by specifying different instance names in your tool calls. The MCP keeps track of all your instances.
What programming languages are supported?
The VPS comes with Node.js pre-installed, but you can deploy any application that can run on Ubuntu Linux. The deployment process is language-agnostic.