mirror of https://github.com/hanzoai/agent.git synced 2026-09-11 21:08:51 +00:00

Files

T

zeekay 2899aeaa9f feat: merge agents control plane + SDKs into unified agent repo

Consolidate all Hanzo agent infrastructure into one repo:
- control-plane/: Go control plane server (was agents-new)
- sdk/: Go, Python, TypeScript SDKs
- deployments/: Docker, Kubernetes, Helm configs
- Dockerfile.sdk: Python SDK image (ghcr.io/hanzoai/agent:latest)
- CI builds both SDK + control-plane images

No more agentfield — everything is hanzoai/agent.

2026-02-13 12:24:31 -08:00

2.6 KiB

Raw Permalink Blame History

Helm (Kubernetes)

This chart installs the Hanzo Agents control plane and optional demo agents.

Quick start (recommended)

Install with PostgreSQL and the demo Python agent (no custom image required):

helm upgrade --install hanzo-agents deployments/helm/hanzo-agents \
  -n hanzo-agents --create-namespace \
  --set postgres.enabled=true \
  --set controlPlane.storage.mode=postgres \
  --set demoPythonAgent.enabled=true

Port-forward the UI/API:

kubectl -n hanzo-agents port-forward svc/hanzo-agents-control-plane 8080:8080

Wait for the demo agent to become ready (first run installs Python deps):

kubectl -n hanzo-agents wait --for=condition=Ready pod -l app.kubernetes.io/component=demo-python-agent --timeout=600s

Open:

http://localhost:8080/ui/

Execute via the control plane:

curl -X POST http://localhost:8080/api/v1/execute/demo-python-agent.hello \
  -H "Content-Type: application/json" \
  -d '{"input":{"name":"World"}}'

Check VCs:

resp=$(curl -s -X POST http://localhost:8080/api/v1/execute/demo-python-agent.hello \
  -H "Content-Type: application/json" \
  -d '{"input":{"name":"VC"}}')
run_id=$(echo "$resp" | python3 -c 'import sys,json; print(json.load(sys.stdin)["run_id"])')
curl -s http://localhost:8080/api/v1/did/workflow/$run_id/vc-chain | head -c 1200

Options

Local storage (SQLite/BoltDB)

helm upgrade --install hanzo-agents deployments/helm/hanzo-agents \
  -n hanzo-agents --create-namespace \
  --set controlPlane.storage.mode=local

Go demo agent (requires a custom image)

The Go demo agent is useful, but you must build/push/load an image that your cluster can pull.

For Minikube, build and load the default image used by the chart:

docker build -t hanzo-agents-demo-go-agent:local -f deployments/docker/Dockerfile.demo-go-agent .
minikube image load hanzo-agents-demo-go-agent:local

helm upgrade --install hanzo-agents deployments/helm/hanzo-agents \
  -n hanzo-agents --create-namespace \
  --set demoAgent.enabled=true

Authentication (optional)

helm upgrade --install hanzo-agents deployments/helm/hanzo-agents \
  -n hanzo-agents --create-namespace \
  --set apiAuth.enabled=true \
  --set apiAuth.apiKey='change-me'

When auth is enabled, API calls must include the key (UI remains accessible):

curl -H "X-API-Key: change-me" http://localhost:8080/api/v1/nodes

Notes

The chart defaults HANZO_AGENTS_CONFIG_FILE=/dev/null so the control plane uses built-in defaults + environment variables.
Admin gRPC listens on (HANZO_AGENTS_PORT + 100) and is exposed via the Service port named grpc.

2.6 KiB Raw Permalink Blame History