Skip to main content
Feedback

AI Gateway Integration

Once your AI Gateway is installed and connected to MCP Gateway, you will begin to see real-time traffic from your MCP servers in the AI Gateway Control Plane.

As MCP Gateway routes requests to your backend tools, the UI will automatically displays:

  • Discovered API endpoints
  • Request performance (latency, error rate, throughput)
  • Live traffic logs by endpoint and by server
  • Gateway routing rules and quota usage

Connection options

You can run MCP Gateway alongside either a hosted AI Gateway or an on-premises deployment.

To enable this, your container must be granted elevated privileges using the flag --cap-add=NET_ADMIN. This grants only your container ability to configure networking, for example, to add firewall rules using iptables.

Hosted gateway

If you are using the Hosted AI Gateway set your API key via the LUNAR_API_KEY environment variable:

interception to hosted gateway
docker run --rm --cap-add=NET_ADMIN -v ./:/lunar/packages/mcpx-server/config -p 9000:9000 -p 5173:5173 -p 9001:9001 -p 3000:3000 -e LUNAR_API_KEY="YOUR_LUNAR_API_KEY" --name mcpx us-central1-docker.pkg.dev/prj-common-442813/mcpx/mcpx:latest

Make sure to replace YOUR_LUNAR_API_KEY with your actual API key found in the AI Gateway Control Plane.

Self-Hosted

If you're running the AI Gateway on your own infrastructure, point MCP Gateway to it using the LUNAR_URL environment variable:

interception for self hosted
docker run --rm --cap-add=NET_ADMIN -v ./:/lunar/packages/mcpx-server/config -p 9000:9000 -p 5173:5173 -p 9001:9001 -p 3000:3000 -e LUNAR_URL="http://localhost:8000" --name mcpx us-central1-docker.pkg.dev/prj-common-442813/mcpx/mcpx:latest

The LUNAR_URL must match the hostname and port used to install the AI Gateway Container.

On this Page