Skip to main content
The ezforge CLI is currently at v0.1.0 (pre-GA). Commands and flags may change before v1.0.0.

​
Prerequisites

  • An ezForge account (sign up free)
  • Docker installed (for building container images)
  • Node.js 20+ or Python 3.11+ (if using a template)

​
Step 1: Install the CLI

brew install ezforgeai/tap/ezforge
Verify the install: 
ezforge --version

​
Step 2: Authenticate

ezforge auth login
This opens your browser to complete the OAuth flow. Your credentials are saved locally and used for all subsequent commands.

​
Step 3: Create a project

ezforge project create my-first-project
Projects are workspaces that group servers, API keys, and billing together.

​
Step 4: Clone the Node.js template

git clone https://github.com/ezforgeai/template-nodejs-mcp-server my-mcp-server
cd my-mcp-server
npm install
The template includes two sample tools out of the box:
  • echo — returns whatever message you send it
  • add — adds two numbers and returns the result

​
Step 5: Deploy

ezforge deploy
ezForge will:
  1. Build your container image
  2. Scan it for vulnerabilities with Trivy
  3. Push it to the registry
  4. Provision a Firecracker microVM
  5. Run a health check at GET /healthz
  6. Swap traffic to the new deployment
Output: 
✓ Building image...
✓ Scanning image (0 critical, 0 high)
✓ Pushing to registry...
✓ Provisioning machine in ord...
✓ Health check passed
✓ Deployed!

Your server is live at: https://my-mcp-server.mcp.ezforge.ai

​
Step 6: Test your server

Connect to your live server from any MCP client: 
https://my-mcp-server.mcp.ezforge.ai/sse
Or test the health check directly: 
curl https://my-mcp-server.mcp.ezforge.ai/healthz
# {"status":"ok"}

​
What’s next?

Add authentication

Protect your server with OAuth 2.1

Node.js template guide

Add your own MCP tools to the template

Python template guide

Deploy a Python-based MCP server

CLI reference

Full CLI command reference