Subcommands

There are several subcommands that can come after “rx”. These take precedence over user commands, e.g., “rx init” will always run the init subcommand, not an “init” binary on the remote host.

help

Print help about the available commands.

Run rx help <subcommand> to see a specific command’s help info.

run

This is the default command if another subcommand isn’t found. It runs a command on the existing remote worker.

$ rx run ls
$ # Equivalent to:
$ rx ls

It is only necessary to explicity use the run subcommand if you’re running a binary with the same name as another rx command. For example:

$ rx run version  # Runs a binary called "version" on the remote

stop

This will stop the current machine you’re using. This stores your current work so you can pick it up on a new machine by simply running any command from your workspace.

For example, suppose you create a file with rx touch foo and then shut down the worker with rx stop. You can then run any command (e.g., rx ls) and rx will provision a new worker with your saved state:

$ rx ls
Error syncing code to your worker, checking with the scheduler...
Your workspace was stowed, please stand by while it is set up on a new machine.
.......
Done! Please rerun this command to use your newly restored workspace.
$ rx ls
foo

subscribe

This will open a browser window and take you through creating a subscription. Subscriptions are charged monthly with a base rate for unlimited CPU usage and an hourly rate for GPU usage. You can cancel anytime from the command line (see the unsubscribe command, below).

See the pricing page page for more info on subscriptions.

unsubscribe

This will cancel your subscription and delete all of your data from rx’s machines. You will receive a final bill via email with any outstanding usage in the current month.

version

Prints the current version of the client you’re using.


Table of contents