Install
Helm chart (recommended)
Section titled “Helm chart (recommended)”Each release publishes the chart as an OCI artifact to GHCR. The chart
is stamped with the matching manager image at release time, so no extra
--set flags are required.
helm install genkit-operator \ oci://ghcr.io/xavidop/charts/genkit-operator \ --version 0.4.1 \ --namespace genkit-operator-system --create-namespaceOmit --version to let Helm pick the highest version available in the
registry, or pin a specific older release by passing its number.
Upgrade in place:
helm upgrade genkit-operator \ oci://ghcr.io/xavidop/charts/genkit-operator \ --version 0.4.1 \ --namespace genkit-operator-systemUninstall:
helm uninstall genkit-operator --namespace genkit-operator-systemYAML bundle (install.yaml)
Section titled “YAML bundle (install.yaml)”A consolidated install file is attached to every GitHub Release.
# Always grabs the latest published release:kubectl apply -f https://github.com/xavidop/genkit-operator/releases/latest/download/install.yaml
# Or pin a specific version:kubectl apply -f https://github.com/xavidop/genkit-operator/releases/download/v0.4.1/install.yamlThis file is generated by make build-installer IMG=… and contains
every CRD, the controller Deployment, the Service, RBAC, and the
namespace — everything needed for a one-shot install.
From source
Section titled “From source”Useful when you want to test an unreleased change:
git clone https://github.com/xavidop/genkit-operatorcd genkit-operatormake install # CRDs onlymake deploy IMG=ghcr.io/xavidop/genkit-operator:v0.4.1Or against a local kind cluster — this builds both images locally, loads them into the cluster, and installs everything in one shot:
kind create cluster --name genkitmake kind-deploy IMG=genkit-operator:devPublished images
Section titled “Published images”| Image | Pull |
|---|---|
| Manager | ghcr.io/xavidop/genkit-operator:<tag> |
| Runner (default) | ghcr.io/xavidop/genkit-runner:<tag> |
| Helm chart | oci://ghcr.io/xavidop/charts/genkit-operator |
Both container images are multi-arch (linux/amd64 and linux/arm64)
and are built as separate per-platform jobs in CI, then merged into a
single manifest list. See the
release reference for details.