Quick Start

Set up MCPX with Docker

1

Create configuration file

Create your config file in your desired folder where your app.yaml and mcp.json will be saved to:

mkdir config
cd config

Once you have completed set up, you can populate your app.yaml and mcp.json using the UI.

2

Run Docker command

In your the config file you just created, run MCPX with the following configuration, exposing ports 9000, 3000 and 5173:

Basic Usage

docker run --rm -v ./:/lunar/packages/mcpx-server/config -p 9000:9000 -p 5173:5173 -p 9001:9001 -p 3000:3000 --name mcpx us-central1-docker.pkg.dev/prj-common-442813/mcpx/mcpx:latest

Without --privileged access MCPX will not be able to run Dockerized MCP Servers:

Advanced allowing Docker-in-Docker (Recommended)

docker run --rm --privileged -v ./:/lunar/packages/mcpx-server/config -p 9000:9000 -p 5173:5173 -p 9001:9001 -p 3000:3000 --name mcpx us-central1-docker.pkg.dev/prj-common-442813/mcpx/mcpx:latest

Exposing port 3000 is optional and used for Prometheus metrics scraping.

Congratulations!

You have now set up your MCPX. Access the MCPX Control Plane by navigating to http://localhost:5173/ in your preferred browser.

Here you can now create your app.yaml and mcp.json configuration files.

What's Next?

Connect your MCP Servers

Connect your MCP servers to MCPX and connect AI agents to MCPX

Connect Agents or AI Applications

Connect to http://localhost:9000/mcp (for Streamable HTTP) or http://localhost:9000/sse (for SSE)