mcp-launcher

🚀 Deploy MCP (Model Context Protocol) servers to Google Cloud Run with a beautiful web dashboard

Features

• 🎛️ Interactive Web Dashboard - Visual interface for managing deployments
• ☁️ One-Click Cloud Run Deployments - Deploy MCP servers in seconds
• 📊 Real-time Logs & Metrics - Monitor your deployments with live updates
• 🔧 Built-in Templates - Start with a working MCP server template
• 🔄 Live Deployment Tracking - Watch your deployment progress in real-time

Prerequisites

Before using mcp-launcher, ensure you have:

1. Node.js 18+ installed
2. Google Cloud SDK (gcloud) installed and authenticated
3. Docker installed and running
4. GCP Project with billing enabled
5. Anthropic API Key (optional, for agent/audit features)

GCP Setup

# Install gcloud CLI
# https://cloud.google.com/sdk/docs/install

# Authenticate
gcloud auth login
gcloud config set project YOUR_PROJECT_ID

# Enable required APIs
gcloud services enable run.googleapis.com artifactregistry.googleapis.com cloudbuild.googleapis.com

Installation

Using npx (Recommended)

npx mcp-launcher

Global Installation

npm install -g mcp-launcher
mcp-launcher

Local Development

git clone <repository-url>
cd mcp-launcher
npm install
cd src/frontend && npm install && cd ../..
npm run build
npm start

Usage

Interactive Mode (Recommended)

Simply run the launcher and follow the prompts:

npx mcp-launcher

The CLI will:

1. Check prerequisites (gcloud, docker)
2. Prompt for your Google Cloud Project ID (if not already configured)
3. Optionally ask for your Anthropic API Key
4. Offer to save configuration to a .env file
5. Launch the web dashboard at http://localhost:3000

Command Line Options

# With all options specified
npx mcp-launcher --project my-project-id --api-key sk-ant-...

# Save configuration for future runs
npx mcp-launcher --project my-project-id --save-config

# Run on a different port without opening browser
npx mcp-launcher --port 8080 --no-browser

# Show help
npx mcp-launcher --help

# Show version
npx mcp-launcher --version

Configuration Methods

The launcher looks for configuration in this order:

1. Command line flags (highest priority)
2. Environment variables
3. Existing .env file in current directory
4. Interactive prompts (if not in CI mode)

Configuration File

You can also create a .env file manually:

# Required
GOOGLE_PROJECT_ID=your_gcp_project_id

# Optional (for agent features)
ANTHROPIC_API_KEY=your_anthropic_api_key_here

# Optional (defaults shown)
PORT=3000
CI=false

Deploying Your MCP Server

Once the dashboard is running:

1. Select your project - Use the built-in template or browse to your MCP server directory
2. Configure deployment - Review the GCP project settings
3. Click Deploy - Watch real-time progress in the dashboard
4. Get your URL - Your MCP server will be live on Cloud Run!

Project Structure

mcp-launcher/
├── bin/                  # CLI entry point
├── src/
│   ├── server/          # Express + Socket.io backend
│   ├── frontend/        # React/Vite dashboard
│   ├── orchestrator/    # Deployment pipeline
│   ├── gcp/             # GCP integrations
│   ├── agent/           # Agent/audit functionality
│   └── config/          # Configuration management
├── template/            # MCP server template
│   ├── math_bot.py      # Example FastMCP server
│   ├── Dockerfile
│   ├── Makefile
│   └── pyproject.toml
├── public/              # Built frontend assets
└── dist/                # Compiled TypeScript

Configuration

CLI Flags

| Flag | Short | Description | |------|-------|-------------| | --project <id> | -p | Google Cloud Project ID | | --api-key <key> | -k | Anthropic API Key (optional) | | --port <number> | | Server port (default: 3000) | | --no-browser | | Don't open browser automatically | | --save-config | -s | Save configuration to .env file | | --ci | | Run in non-interactive mode | | --help | -h | Show help message | | --version | -v | Show version number |

Environment Variables

| Variable | Required | Description | |----------|----------|-------------| | GOOGLE_PROJECT_ID | Yes | Your GCP project ID | | ANTHROPIC_API_KEY | No | Your Anthropic API key for agent features | | PORT | No | Server port (default: 3000) | | CI | No | Set to true to disable browser auto-open |

Frontend Development

cd src/frontend
npm install
npm run dev        # Vite dev server on port 5173
npm run build      # Build for production

Deployment Process

The launcher automates these steps:

1. GCP Setup - Enables required APIs and creates Artifact Registry
2. Docker Build - Builds container image for your MCP server
3. Push to Registry - Pushes image to Google Artifact Registry
4. Cloud Run Deploy - Deploys to Cloud Run with auto-scaling
5. Verification - Confirms deployment and captures service URL

Template

The included template (template/) provides a working MCP server:

• FastMCP - Modern Python MCP framework
• HTTP Transport - Ready for Cloud Run
• Tools - Example add tool
• Resources - Dynamic greeting resources
• Prompts - Configurable greeting prompts

Customize template/math_bot.py to create your own MCP server!

Troubleshooting

Docker not running

# macOS
open -a Docker

# Linux
sudo systemctl start docker

gcloud not authenticated

gcloud auth login
gcloud auth configure-docker

Permission denied

# Ensure you have the right GCP permissions
# Required roles: Cloud Run Admin, Artifact Registry Admin, Service Account User

Development

# Install dependencies
npm install
cd src/frontend && npm install

# Build everything
npm run build

# Run in development mode (backend only)
npm run dev

# Run frontend dev server
cd src/frontend && npm run dev

License

ISC

Contributing

Contributions welcome! Please read our contributing guidelines and submit PRs.