Getting Started

Local development

The fastest way to try arkonis-operator is with the make dev command. It creates a kind cluster and sets everything up in one step with no manual Redis config required.

Prerequisites: Docker, kind, kubectl, Go 1.25+

git clone https://github.com/arkonis-dev/arkonis-operator.git
cd arkonis-operator
make dev ANTHROPIC_API_KEY=sk-ant-...

This will:

  1. Create a kind cluster
  2. Build and load both Docker images into the cluster
  3. Install the CRDs
  4. Deploy Redis and the operator inside the cluster
  5. Create the API key secret

When it finishes, deploy your first agent:

kubectl apply -f config/samples/arkonis_v1alpha1_arkonisdeployment.yaml
kubectl get aodep -w
# NAME             MODEL                      REPLICAS   READY   AGE
# research-agent   claude-sonnet-4-20250514   2          2       30s

Tear down when done:

make dev-down

Production install

Prerequisites

  • Kubernetes 1.31+
  • kubectl configured against your cluster
  • An API key for your chosen LLM provider (Anthropic by default)

1. Install the operator 

kubectl apply -f https://github.com/arkonis-dev/arkonis-operator/releases/latest/download/install.yaml

This installs the CRDs, RBAC, and the operator deployment in the arkonis-system namespace.

2. Deploy Redis

arkonis-operator uses Redis Streams as the task queue between the operator and agent pods.

kubectl apply -f https://raw.githubusercontent.com/arkonis-dev/arkonis-operator/main/config/prereqs/redis.yaml

3. Create the API key secret

Create one secret per namespace where agents will run. The operator injects the entire secret as environment variables into every agent pod. Add whichever keys your chosen provider requires.

Anthropic (default):

kubectl create secret generic arkonis-operator-api-keys \
  --from-literal=ANTHROPIC_API_KEY=sk-ant-... \
  --from-literal=TASK_QUEUE_URL=redis.agent-infra.svc.cluster.local:6379

TASK_QUEUE_URL is always required. To use a different provider, also add AGENT_PROVIDER=<name> to the secret.

4. Deploy your first agent 

kubectl apply -f https://raw.githubusercontent.com/arkonis-dev/arkonis-operator/main/config/samples/arkonis_v1alpha1_arkonisdeployment.yaml

Verify 

kubectl get aodep
# NAME              MODEL                      REPLICAS   READY   AGE
# research-agent    claude-sonnet-4-20250514   2          2       30s

kubectl get pods -l arkonis.dev/deployment=research-agent
# NAME                              READY   STATUS    RESTARTS   AGE
# research-agent-agent-7d9f-xk2p8   1/1     Running   0          30s
# research-agent-agent-7d9f-mn4q1   1/1     Running   0          30s

Next steps

Apache 2.0 · ARKONIS