OpenFrame Documentation

Quick Start

OpenFrame CLI is a Go command-line tool that provisions a Kubernetes cluster (local k3d, or cloud GKE/EKS via Terraform) and deploys the OpenFrame platform onto it through ArgoCD's app-of-apps pattern. This page gets you from zero to a running OpenFrame platform in a single command.

Note on scope: OpenFrame CLI (flamingo-stack/openframe-cli) drives the deployment process. The platform it installs — the services, charts, and application code that end up running on your cluster — lives in the separate openframe-oss-tenant repository (https://github.com/flamingo-stack/openframe-oss-tenant). This quick start only covers installing and running the CLI itself.

Hardware requirements

Before you start, make sure your machine (or the target cluster) meets these requirements:

The openframe app install command also warns (without blocking) if the target cluster reports less than roughly 6 CPU cores / 24GB RAM.

TL;DR: 5-minute setup

1. Install the CLI

Windows:

Download the latest release archive and run the extracted binary:

https://github.com/flamingo-stack/openframe-cli/releases/latest/download/openframe-cli_windows_amd64.zip

macOS / Linux:

Grab the matching archive for your platform from the releases page:

https://github.com/flamingo-stack/openframe-cli/releases

Each release ships signed binaries: macOS binaries are codesigned and notarized, and Windows binaries are Authenticode-signed (see scripts/sign-binary.sh). Extract the archive and place the openframe binary on your $PATH.

Once installed, you can always check for a newer release and apply it in place:

openframe update check
openframe update

2. Run bootstrap

A single command creates a local k3d cluster and installs the full OpenFrame platform on it:

openframe bootstrap

This is equivalent to running openframe cluster create followed by openframe app install, but streamlined into one interactive flow (see cmd/bootstrap/bootstrap.go and internal/bootstrap/service.go).

# Give the cluster a specific name
openframe bootstrap my-cluster

# Non-interactive (CI/CD): reuses an existing openframe-helm-values.yaml
openframe bootstrap --non-interactive

# See everything happening under the hood, including ArgoCD sync progress
openframe bootstrap --verbose

3. Check status

openframe app status

That's it — you have a cluster with ArgoCD and the OpenFrame app-of-apps running.

"Hello World": step by step

If you'd rather see each stage explicitly instead of running bootstrap, here is the same flow broken into its two underlying commands.

Step 1 — Create a cluster

openframe cluster create

Run without flags, this opens an interactive wizard where you choose a cluster type (local k3d, or cloud GKE/EKS). Add --skip-wizard to create a cluster directly from flags and defaults, useful for scripting:

openframe cluster create --skip-wizard

The cluster command group also supports list, status, use, delete, and cleanup (see cmd/cluster/cluster.go). For local development, k3d is the fastest path since it runs entirely on Docker with no cloud billing involved.

Step 2 — Install the OpenFrame platform

Once a cluster exists and is reachable, install ArgoCD and the app-of-apps onto it:

openframe app install

Run interactively (no arguments), this lets you pick which kube-context to target. You can also target a specific cluster directly:

openframe app install my-cluster
openframe app install --context k3d-my-cluster

Under the hood, app install (cmd/app/install.go) resolves a rest.Config for the target cluster, then calls services.InstallChartsWithConfigContext, which installs ArgoCD with pinned chart values and then the app-of-apps from the configured GitHub repository/ref (chartmodels.RepoOSSTenant by default, overridable with --ref).

# Deploy a specific release tag instead of the default branch
openframe app install --ref 1.0.48

Step 3 — Watch it come up

openframe app status --watch

--watch re-renders the status every few seconds until the platform's ArgoCD applications are Synced and Healthy (cmd/app/status.go). Press Ctrl+C to stop watching.

Expected output

openframe bootstrap

A successful bootstrap prints a summary card listing the stages completed (cluster creation, ArgoCD install, app-of-apps sync), their timings, and access hints for the platform once everything is ready. Internally this follows the sequence described in docs/diagrams/architecture/bootstrap-sequence.mmd: cluster creation → ArgoCD Helm install → app-of-apps clone and install → wait for every ArgoCD Application to report Synced/Healthy.

openframe app status

Cluster reachable (3/3 nodes ready)
APPLICATION          SYNC      HEALTH
argocd               Synced    Healthy
app-of-apps          Synced    Healthy
...
All applications are Synced and Healthy.

If the platform is not yet installed, you'll instead see:

No OpenFrame applications found — is it installed? Run: openframe app install

You can also request machine-readable output for scripting:

openframe app status --output json
openframe app status --output yaml

And drive an interactive terminal dashboard (k9s-style) to navigate applications and trigger syncs:

openframe app status --interactive

openframe app access

Once the platform is installed, retrieve the ArgoCD sign-in details:

openframe app access

Expected output:

ArgoCD access
  Username: admin
  Password: <redacted>
Open the ArgoCD UI:
  1. kubectl port-forward -n argocd svc/argocd-server 8080:443
  2. open https://localhost:8080

Command reference at a glance

A full architecture walkthrough — including the command-to-service dependency diagram and the app-status aggregation sequence — is available in docs/reference/architecture/overview.md and docs/diagrams/architecture/README.md.

How this fits together

flowchart LR
  CLI["OpenFrame CLI"]
  Cluster["cluster create / k3d, GKE, EKS"]
  ArgoCD["ArgoCD install"]
  AppOfApps["app-of-apps (openframe-oss-tenant)"]
  K8s[("Kubernetes API")]

  CLI --> Cluster
  Cluster --> K8s
  CLI --> ArgoCD
  ArgoCD --> K8s
  ArgoCD --> AppOfApps
  AppOfApps --> K8s

The CLI itself has no recorded upstream dependencies on other internal repositories and no recorded downstream consumers in the organization's dependency graph — it is a standalone, leaf artifact published as the Go module github.com/flamingo-stack/openframe-cli. The platform it deploys (charts, services, application code) is sourced from the separate openframe-oss-tenant repository at install time, via the git ref you pass with --ref (default branch otherwise).

Next steps

Now that you have a cluster and the platform running:

• Work through the detailed how-to guides for each stage of the lifecycle: creating a cluster, installing the platform, checking status, updating the CLI, and uninstalling — see Create a Kubernetes cluster with OpenFrame CLI, Install the OpenFrame platform onto a cluster, Check the status of your OpenFrame cluster and app, Update the OpenFrame CLI to the latest version, and Uninstall OpenFrame and delete a cluster.
• Read the Introduction and Prerequisites pages for background and tool requirements before diving deeper.
• Continue to First Steps for a guided walkthrough beyond the initial bootstrap.
• For contributors, see the Development overview, environment setup, local development guide, architecture docs, security notes, testing guide, and contributing guidelines.
• For a deeper architectural reference, see Architecture overview, the ecosystem graph, and the architecture diagrams.

For community support and discussion, OpenFrame does not use GitHub Issues or Discussions — join the OpenMSP Slack community at https://www.openmsp.ai/.