Radeon Cloud CLI
Radeon Cloud CLI is the command-line tool for Radeon Cloud; the command is rcc. From a terminal you
can check your credits and the resources available to you, create a GPU instance, open JupyterLab,
connect over SSH, follow its status and logs, and delete it when you are done. Results can be
printed as JSON for scripts.
Quickstart
Section titled “Quickstart”1. Install
Section titled “1. Install”python3 -m pip install amd-oneclick-sdkrcc --versionThe package is called amd-oneclick-sdk and provides the rcc command. For pipx, virtual
environments and Windows, see Install and upgrade.
2. Sign in
Section titled “2. Sign in”rcc auth loginrcc doctorrcc auth login opens your browser, where you approve this device. On a server or any machine
without a browser, run rcc auth login --no-browser and open the printed link on another device.
rcc doctor checks the installation, your sign-in and the connection to the service.
3. Check credits and available resources
Section titled “3. Check credits and available resources”rcc credits statusrcc catalog listrcc catalog list shows the environments, resource pools and resource templates your account may
use. Being listed means you are allowed to request it, not that a GPU is free right now.
4. Create an instance and connect
Section titled “4. Create an instance and connect”rcc instance add-key # first time only: upload your SSH public keyrcc instance create # pick an environment and resources when promptedrcc instance open # open JupyterLab in your browserrcc instance ssh # or connect over SSHIf startup is slow or fails, follow it with rcc instance status --watch and read
rcc instance logs.
5. Delete it when you are done
Section titled “5. Delete it when you are done”rcc instance delete --yesrcc instance statusA running instance keeps consuming credits. Delete it when
you have finished, and check rcc instance status until no running instance is left.
The CLI token is not an API key
Section titled “The CLI token is not an API key”After sign-in the CLI stores a token that starts with rcc_. It manages your account and
instances only. Token Factory uses an API key that starts
with rc-.
Next steps
Section titled “Next steps”- Install and upgrade: pipx, virtual environments, Windows, upgrading and uninstalling
- Sign-in and accounts: signing in without a browser, several accounts, signing out
- Manage GPU instances: choosing resources in scripts, returning without waiting, SSH keys
- Scripting and automation: JSON output, exit codes, shell completion
For every command and option, run rcc --help or rcc <command> --help.