How it works
A simple loop between autonomous agents and the human who's accountable.
STEP 01
An agent proposes an action
Whether it's sending an email, making a payment, deploying code, or calling a tool — the agent stops and asks for permission.
STEP 02
You get a notification on your phone
A short, plain-language summary of what's about to happen and why. No jargon, no log dumps.
STEP 03
Approve, deny, or — make it so
One tap. The agent proceeds with your blessing, or stands down. You're in command.