OpenFrame Documentation
Deployment guides, the CLI reference, architecture and admin guides for OpenFrame
Flamingo Raises $4.5M Seed Round · Led by Vertex Ventures, bringing total funding to $6.7M to build the open, AI-native layer for IT and security.
Deployment guides, the CLI reference, architecture and admin guides for OpenFrame
This page walks you through what to do right after installing the OpenFrame CLI (openframe): the first commands to run, the configuration you'll typically touch, the key features worth exploring, and where to go if you get stuck.
If you have not installed the CLI yet, see ./introduction.md and ./prerequisites.md first. If you want the fastest path to a running platform, ./quick-start.md covers the single-command bootstrap flow — this page instead walks through the individual steps and what each one configures.
Scope reminder: the OpenFrame CLI (this repository) is the tool that provisions clusters and deploys the platform. The platform's application code — the services that actually make up OpenFrame — lives in the separate
flamingo-stack/openframe-oss-tenantrepository, which the CLI clones and deploys for you via its app-of-apps Helm chart. You won't need to clone that repository yourself for the steps below.
OpenFrame deploys a full Kubernetes-based platform (ArgoCD, the app-of-apps chart, and the services it manages), so the machine running the cluster needs real headroom:
| Resource | Minimum | Recommended |
|---|---|---|
| RAM | 24 GB | 32 GB |
| CPU Cores | 6 | 12 |
| Disk Space | 50 GB | 100 GB |
Run the root command with no arguments, or check the version metadata that cmd/root.go resolves at build time:
openframe --help
This prints the command tree: bootstrap, cluster, app, prerequisites, and update. If this prints help text instead of a "command not found" error, the install succeeded.
Before provisioning anything, verify your machine has the tools OpenFrame needs (Docker, k3d, Helm, and — for cloud clusters — Terraform, gcloud, or the AWS CLI):
openframe prerequisites check
This reports which required tools are present without changing anything. If something is missing:
openframe prerequisites install
On macOS and Linux this auto-installs missing tools. On Windows it prints manual installation links instead of installing automatically. You can scope either command to a specific cluster backend with --type k3d, --type eks, or --type gke.
Some of the tools this step installs — k3d, Helm, mkcert, Terraform, and infracost — are downloaded by the CLI itself through its internal download pipeline, pinned to exact versions and verified against SHA256 digests before use, never fetched via an unverified pipe-to-shell.
With prerequisites satisfied, create a cluster. For a local cluster (the fastest option, backed by k3d/Docker):
openframe cluster create
Run interactively (it will prompt you for a cluster name and options), or pass a name directly and skip the wizard:
openframe cluster create my-cluster --skip-wizard
For cloud clusters, pass --type eks or --type gke; cloud creation shows a Terraform plan (and an optional infracost estimate) before applying anything. A full, detailed walkthrough of this step — including the wizard options and non-interactive flags — continues in ./how-to/create-a-new-kubernetes-cluster-with-openframe-cli.md.
If you'd rather combine cluster creation and platform installation into one command,
openframe bootstrap [cluster-name]does both steps in sequence (with--non-interactivefor CI, and--verbosefor detailed ArgoCD sync progress) — see the bootstrap sequence in../reference/architecture/overview.md.
Once you have a reachable cluster (created with openframe cluster create, or any other cluster you point --context at), deploy ArgoCD and the app-of-apps chart:
openframe app install
or, targeting a specific cluster by name:
openframe app install my-cluster
This is where the CLI clones the flamingo-stack/openframe-oss-tenant repository (via the GitHubRepo/GitHubBranch settings in internal/chart/models/appofapps.go) and installs it through Helm and ArgoCD. Pass --ref to deploy a specific branch or release tag instead of the default. The full procedure, including configuration prompts, continues in ./how-to/install-the-openframe-platform-onto-a-cluster.md.
After installation finishes (or while it's in progress), check status:
openframe app status
This reports cluster reachability plus ArgoCD application sync/health. Add --watch to keep it refreshing, or --interactive to open a k9s-style terminal UI for browsing application state live. The detailed status workflow continues in ./how-to/check-the-status-of-your-openframe-cluster-and-apps.md.
Once the platform is healthy, get the ArgoCD admin credentials and the port-forward command to reach its UI:
openframe app access
Non-default installs are driven by a Helm values file named openframe-helm-values.yaml, read from and written to your current working directory (see internal/chart/utils/config/paths.go). When you run openframe app install interactively, the configuration wizard (internal/chart/ui/configuration/wizard.go) offers two modes:
openframe-helm-values.yaml as-is, or creates one with defaults if none exists.Configuration Mode
> Default configuration
Interactive configuration
Controls which branch of flamingo-stack/openframe-oss-tenant gets deployed (internal/chart/ui/configuration/branch.go). You can keep the current branch or specify a custom one — useful for testing a feature branch of the platform before it merges.
Lets you supply Docker registry credentials (username, password/token, email) if your deployment needs to pull from a private registry (internal/chart/ui/configuration/docker.go). Credentials you enter here are registered with the CLI's secret redactor so they never appear in verbose logs or error output.
Chooses how the platform is exposed (internal/chart/ui/configuration/ingress.go):
API keys, auth tokens, and Docker passwords entered during the wizard are registered as secrets immediately at the point of collection — before they can reach any log output.
Commands under app (install, upgrade, status, access, uninstall) take no positional arguments for targeting — use --context explicitly instead, to avoid accidentally operating against whatever cluster is the current kube-context. This is enforced deliberately: passing a stray cluster name to a destructive command like uninstall used to run silently against the current context.
If you manage multiple clusters, switch which one kubectl/the CLI targets:
openframe cluster use my-cluster
For GKE clusters this also updates your local gcloud configuration.
The cluster command group covers the full lifecycle beyond creation:
openframe cluster create|delete|list|status|use|cleanup
list [--all] — lists clusters the CLI manages; --all additionally discovers external GKE/EKS clusters read-only.status [NAME] — detailed health, nodes, and installed applications, with -o json|yaml for scripting.cleanup [NAME] — prunes unused container images from cluster nodes without touching installed apps.delete [NAME] — destroys a cluster; cloud deletes require typed name confirmation unless --force.Two distinct upgrade modes are available under openframe app upgrade:
--ref (a different branch or tag of the platform).--sync (optionally with --prune).The full upgrade procedure, including when to use each mode, continues in ./how-to/upgrade-openframe-cli-and-its-installed-apps.md.
The CLI can update itself:
openframe update [version]|check|rollback
update [version] — updates to the latest release, or switches to a specific tagged version (up or down). Every download is checksum- and signature-verified via Sigstore/cosign, anchored through TUF to the public-good Sigstore trust root.update check — reports update availability without changing anything.update rollback — reverts to the previously-installed version offline, with no download.When you're done with a cluster, or want to remove just the platform while keeping the cluster running:
openframe app uninstall
removes ArgoCD and the app-of-apps Helm releases/Applications, in a safe order (Applications deleted first, then Helm releases, then finalizers cleared), leaving the cluster itself intact.
To remove the cluster entirely as well, continue to ./how-to/uninstall-openframe-and-delete-a-cluster.md.
On Windows, cluster and platform operations transparently re-execute inside WSL2 before running, because the local cluster (Docker + k3d) and the native Kubernetes client need to share a network namespace that a native Windows process can't reach reliably. You don't need to do anything differently — just run openframe commands as usual from Windows; the WSL launcher handles the handoff and auto-installs the Linux binary into WSL if it isn't there yet.
For a deeper look at how these commands map to the CLI's internal services (cluster provisioning, chart/ArgoCD installation, prerequisites, self-update) and how they talk to Docker, Kubernetes, cloud provider APIs, and GitHub Releases, see ../reference/architecture/overview.md and ../reference/architecture/ecosystem.md. Diagram sources are also available under ../diagrams/architecture/README.md.
Per the organization's repository graph, OpenFrame CLI publishes a single Go module (github.com/flamingo-stack/openframe-cli) and has no recorded upstream dependencies or downstream consumers within the organization — it's a standalone, user-facing binary distributed via GitHub Releases and go install, not a library other services import.
This project does not use GitHub Issues or GitHub Discussions. All community support, questions, and discussion happen on the OpenMSP Slack community:
For background on the wider Flamingo platform and OpenFrame product, see https://flamingo.run and https://openframe.ai.
If you want to go deeper on contributing to the CLI itself, see ../development/README.md, the environment and local-development setup guides at ../development/setup/environment.md and ../development/setup/local-development.md, the architecture notes at ../development/architecture/README.md, the security guidance at ../development/security/README.md, the testing guide at ../development/testing/README.md, and the contribution guidelines at ../development/contributing/guidelines.md.
Remember: the platform application code you're deploying lives in
flamingo-stack/openframe-oss-tenant, not in this CLI repository — bug reports or questions about the deployed services themselves ultimately trace back there, but for CLI issues and general platform questions the OpenMSP Slack is the shared channel either way.