v0.1.0 · open source · alpha

Stop deploying OpenShift the hard way.

An AI copilot that walks you through AWS IPI deployments, generates install-config.yaml from a description, and asks for your approval before running anything destructive. Plain English in, deployments out.

pip install cloudnest-pilot
View on GitHub → See it in action
Apache 2.0 licensed Bring your own API key Alpha — actively developed
localhost:8765 — Cloudnest Pilot
Cloudnest Pilot showing an approval card before running a shell command

Built for engineers who'd rather not remember every IAM permission.

Three things you'll actually feel when you use it.

⌨

Plain English in

"Generate an install-config for OpenShift 4.15 in us-east-1, 3 masters, 2 workers." That's the prompt. No `openshift-install` flag-hunting, no copy-pasting from Confluence pages.

✓

Approval-gated execution

Every shell command shows you exactly what will run before it runs. A hard-coded blocklist refuses obviously destructive patterns even if you click approve. Worst case: you deny something and try again.

⌘

Open source, BYO key

Apache 2.0. Your Anthropic API key talks to Anthropic directly — we never proxy your conversations. Your AWS credentials never leave your machine. Audit log written to ~/.cloudnest-pilot/history.log.

Try it in two commands.

Demo mode runs without an API key — useful for showing colleagues what the tool does before they invest in setup.

  1. Install from PyPI

    Requires Python 3.10+. We recommend a fresh virtualenv.

    pip install cloudnest-pilot

  2. Try demo mode (no API key needed)

    Opens a chat UI at http://localhost:8765 with canned responses. Good for kicking the tires.

    cloudnest-pilot --demo

  3. Add your Anthropic API key for the real thing

    Set it in ~/.cloudnest-pilot/.env as ANTHROPIC_API_KEY=sk-ant-..., then:

    cloudnest-pilot --web
🛡

Honest framing. This is v0.1.0 alpha — it works in our testing, but expect rough edges. The approval gate ensures the worst case is "deny something unexpected and try again." We mark this Stable only after dozens of real, successful production deployments.