Self-hosting

The platform is designed to be self-hosted. There is no required hosted control plane. The landing site at www.lamboapp.com is a convenience for signup + fork provisioning; you can skip it entirely and clone the upstream repo directly if you'd rather.

Local development (kind)

The default configuration boots on kind, macOS or Linux. Cost: $0.

brew install kind kubectl helm
git clone https://github.com/jaredzwick/equity-platform ~/equity-platform
cd ~/equity-platform
./local/up.sh

That's it. ./local/up.sh is idempotent — run it as many times as you like.

Production (any conformant Kubernetes)

The manifests under bootstrap/, apps/, and charts/ are cluster-portable. To adopt them on a production cluster:

Point ArgoCD at your fork. Configure argocd-cm with your repo URL and a deploy key or GitHub App credential.
Apply the root app-of-apps — kubectl apply -f bootstrap/root-app.yaml. ArgoCD picks up everything from there.
Swap the External Secrets backend if you're not using local secrets. ESO supports DO Secrets, 1Password, AWS Secrets Manager, GCP Secret Manager, HashiCorp Vault — the ClusterSecretStore is a one-file swap under bootstrap/.
Point DNS at your ingress. envoy-gateway is the default; swap for your ingress of choice if needed.

A DO Kubernetes overlay is on the roadmap. Until then, treat the kind config as the reference — the delta for a managed k8s provider is mostly ingress + secrets.

Environment variables

Full template lives at console/.env.example in the repo. The important ones:

Variable	Required for	Notes
`KUBECONFIG`	Any k8s reads	Points to your cluster. Defaults to `~/.kube/config`.
`GITHUB_TOKEN`	Write-back (new app, new business)	Fine-grained PAT with Contents: Read and Write scoped to your fork.
`GITHUB_REPO`	Write-back	Format `owner/repo`, e.g. `yourname/equity-platform`.
`NATS_MONITOR_URL`	Events tab	Defaults to `http://localhost:8222` — port-forward from the cluster.

The console never sends anything to a hosted control plane. All state is either in your git repo, your cluster, or your local disk.

Why we don't offer a fully hosted platform

The console holds live connections to your Kubernetes API. Vercel-style serverless can't do that — no persistent process, no persistent connection. A fully hosted control plane would need dedicated infra (a bastion, a websocket relay, a per-customer proxy), and the cost story stops being "$0 to try."

If you want a hosted control plane, get in touch (commercial@pypes.dev). Otherwise: self-host, and pay only for the cluster you're running your businesses on anyway.