kagent

Deploy your agentregistry MCP server to a Kubernetes cluster.

Before you begin

Follow the Get started guide to install agentregistry.
Connect the kagent runtime so that you can deploy MCP servers to your Kubernetes cluster.
Publish an MCP server.
If you are using a kind cluster and pushed the image to a local registry rather than a remote one, load the image into the cluster. Replace ghcr.io/my-org/mymcp:latest with the image identifier from your mcp.yaml file and <cluster-name> with the name of your kind cluster.
```
kind load docker-image ghcr.io/my-org/mymcp:latest --name <cluster-name>
```

Deploy the MCP server

List the runtimes available in agentregistry. Verify that you see the kubernetes-default runtime, which is the runtime that you use to deploy MCP servers to Kubernetes by using the kagent OSS project.
```
arctl get runtimes
```
Example output:
```
NAME                 TYPE
kubernetes-default   Kubernetes
local                Local
```

Create a Deployment resource that deploys a published MCP server to the kubernetes-default runtime.

arctl apply -f- <<EOF
apiVersion: ar.dev/v1alpha1
kind: Deployment
metadata:
  name: mymcp-server
spec:
  env:
    KAGENT_NAMESPACE: default
  runtimeRef:
    kind: Runtime
    name: kubernetes-default
  targetRef:
    kind: MCPServer
    name: mymcp
    tag: latest
EOF

Example output:

✓ Deployment/mymcp-server created

List the deployments and verify that mymcp-server appears.

arctl get deployments

Example output:

NAME                    TARGET   VERSION   TYPE   RUNTIME              STATUS
default/mymcp-server   mymcp    latest    mcp    kubernetes-default   deploying

Verify the deployment

Verify that kagent created an MCPServer resource in your cluster.

kubectl get mcpservers -o yaml

Example output:

apiVersion: v1
items:
- apiVersion: kagent.dev/v1alpha1
  kind: MCPServer
  metadata:
    annotations:
      aregistry.ai/deployment-id: my-mcp-server
    creationTimestamp: "2026-08-19T21:10:05Z"
    generation: 1
    labels:
      aregistry.ai/deployment-id: my-mcp-server
      aregistry.ai/managed: "true"
    name: mymcp-my-mcp-server
    namespace: default
    resourceVersion: "103614"
    uid: 26fb8faa-3279-4aea-b8be-b504f2125479
...

Verify that a pod was created for your MCP server and is running in your cluster.

kubectl get pod | grep mymcp

Example output:

mymcp-my-mcp-server-77694d689d-gv98q   1/1     Running   0          2m37s

Test the deployed server

Port-forward the MCP server service.

kubectl port-forward service/mymcp-mymcp-server 3000

In a separate terminal, list the available tools. Verify that you see the example_echo and example_sum tools.

mcp-inspector --cli http://localhost:3000/mcp \
  --transport http --method tools/list | jq '.tools[].name'

Example output:

"example_echo"
"example_sum"

Call the example_echo tool and verify that you get back the echoed message.

mcp-inspector --cli http://localhost:3000/mcp \
  --transport http --method tools/call \
  --tool-name example_echo \
  --tool-arg message="hello world"

Optional: Open the kagent dashboard and go to View > MCP & Tools to see the my-mcp server and its tools listed.
```
kagent dashboard
```

Cleanup

List the deployments and find the one you want to delete.
```
arctl get deployments
```
Remove the deployment.
```
arctl delete deployment my-mcp-server
```