Planton

Deploy real infrastructure to your own cloud — without writing Terraform.

Planton is a free desktop app and CLI you download and open. It finds the cloud you're already signed into; pick a stack and fill a short form, or planton apply -f a manifest from your terminal — and watch it deploy, with clean, auditable infrastructure-as-code running underneath. No account. No connections. No ceremony.

Website · Getting Started · Documentation

---

How it works

You never start from a blank Terraform file. Planton runs proven, pre-built infrastructure-as-code modules — built for secure, well-architected, cost-efficient defaults — against your own cloud account, with real state and history. Create like a console: pick a stack, fill a short form. Manage like Kubernetes: planton apply -f deploys one manifest (one Infra Component) or a whole directory of them in dependency order, and planton chart install stands up a templated environment — the kubectl apply and helm install gestures, freed from Kubernetes and extended to every cloud.

Whichever way you deploy, it's real infrastructure-as-code the whole way down — stored, versioned, every change a diff. On your own cloud. Nothing to lock you in.

What's in this repository

This repo holds the open building blocks that power Planton, all under Apache-2.0. Audit them, fork them, or take your configuration and run it yourself.

• catalog/ — 700+ catalog kinds across 8 cloud providers (AWS, GCP, Azure, Kubernetes, DigitalOcean, Cloudflare, Auth0, and OpenFGA). Each kind is a Protocol Buffer definition in the Kubernetes Resource Model shape (apiVersion/kind/metadata/spec) with field-level validations and auto-generated SDKs in Go, Python, TypeScript, and Java — plus verified, machine-checked fact-sheets: cost data with generated per-preset estimates (priced from the pinned price books in catalog/_pricing/), technical control posture with framework crosswalks (catalog/_compliance/), and least-privilege provisioning-permission manifests validated against the providers' own published inventories.
• charts/ — 18 ready-made infra charts: whole environments (network + compute + data + DNS) composed from the kinds above and installed in one command — the Helm-chart idea, for cloud infrastructure.
• cmd/planton — the open-source CLI and IaC engine that validates manifests and executes the Pulumi and OpenTofu/Terraform modules that ship with every kind.
• site/ — the source of planton.ai: website, docs, and blog (Next.js static export, deployed to GitHub Pages on every site/ change). Run it locally with make run-site, or build and preview the static export with make preview-site.

The CLI

The desktop app is the product; the CLI is its companion — the same deploys, from your shell, driving the same engine.

brew install plantonhq/tap/planton

Write a manifest in the shape you already know:

apiVersion: kubernetes.planton.dev/v1alpha1
kind: KubernetesPostgres
metadata:
  name: my-first-postgres
spec:
  namespace:
    value: my-first-postgres
  createNamespace: true
  container:
    replicas: 1
    diskSize: 1Gi

Then deploy it:

planton apply -f postgres.yaml

Validation catches mistakes in seconds — before anything touches your cloud — and the deploy streams live output from the underlying IaC module. See the Getting Started guide and the CLI reference.

From your coding agent

Planton ships its assistant's knowledge as agent skills in the open Agent Skills format, so Cursor, Claude Code, Codex, and any other coding agent can compose, validate, price, and deploy from the same craft the platform's own assistant uses. Install them once:

npx skills add plantonhq/skills

Two skills ship together: planton (the working craft, and the boundaries the agent never crosses — no mutation without your consent, never outside your repository) and multi-cloud-catalog (one reference page per kind, with its cost, control posture, and runner permissions). Agents that want the platform's own operations reach them over MCP at https://mcp.planton.ai/. See Coding Agents.

Where it runs

Planton runs where you decide: hosted at planton.ai, self-hosted on your own Kubernetes cluster with a license that verifies offline, or free on your laptop as Planton Desktop. In every shape it is your cloud account, your keys, your state, and your bill. The same manifests and the same model run on all three, so nothing is redone when a person becomes a team.

Self-hosting is two Helm installs — the operator, then the platform at the release you name — with the control plane, the console, the identity server, the secrets manager, the databases, and an in-cluster runner all inside your cluster. See Self-Hosting.

Licensing

Planton the app is free, including for commercial use. The building blocks that power it — the infrastructure kinds, the charts, and the CLI — are open source under Apache-2.0: audit them, fork them, or take your configuration and run it yourself. No lock-in. The Planton name and logo are trademarks of Planton Cloud, Inc. — the code is yours to fork, the identity is not; see TRADEMARKS.md.

Contributing

Visit CONTRIBUTING.md for information on building Planton from source and for details about becoming a contributor.

Acknowledgments

• Brian Grant & the Kubernetes API team for their foundational work on the Kubernetes Resource Model.
• The Protobuf Team for laying the foundation for a powerful language-neutral contract definition language.
• The Buf Team for their Protobuf tooling — including BSR Docs, BSR SDKs, and ProtoValidate — which collectively democratized protobuf adoption and made this project possible.
• The Pulumi team for providing a powerful infrastructure-as-code platform that enables multi-language support.
• The spf13/cobra team for making building command line tools a bliss.

---