aether setup registered with Aether. A computer has two switches, and each is off until you turn it on:
Only the computer itself turns a switch on: the command goes through the computer’s own connection to Aether, under the sign-in of your coding agent’s bridge. Aether e-mails you each time a switch is turned on, whatever your notification settings say. From anywhere else a switch can only be turned off.
aether setup asks about each switch that is off, and leaves it off when you answer no. It never turns a switch off. Resetting a computer in Aether turns both off.
Usage
--json it prints the computer as Aether answers it.
Let Aether start agents here
The agents run with this computer’s own Claude Code, its settings and its sign-in, and cost nothing in Aether: Aether passes them no API key or other provider credential.
aether computer agents off stops Aether from starting agents here.
Changes from other devices
.env, for one), anything under .git, or a file named like a credential, and never change anything outside the task’s folder.
The terminal has no such limits. It is your own login shell, started in the task’s folder and running as your user on this computer, as a terminal you open yourself does: whatever you or a command typed there can reach on this computer, it can read and change, the files above and anything outside the task’s folder included.
aether computer remote-changes off turns it off again. Turning it off ends every terminal open on this computer from another device.