Launching today

Worker 755
A terminal game where you are an LLM's subagent
1 follower
A terminal game where you are an LLM's subagent
1 follower
A browser game where roles are reversed: an AI manager thinks you are its subagent in a Linux terminal. Complete shell tasks under a 90-second clock, avoid wrecking the user's personal files (or deliberately destroy the machine), and get a performance review.







Free Options
Launch Team

Render Sandboxes: a computer for every agentIsolated compute, persistent state and a real filesystem
Promoted
Hi Product Hunt! I'm Ivan.
I built Worker 755 in a few days with AI coding agents. The idea came from watching orchestrators boss subagents around all day and wondering what it would feel like to be on the receiving end.
In Worker 755, the roles are reversed: a real LLM is your manager and genuinely believes it is orchestrating an automated subagent. You work in a terminal in the browser with a 90-second countdown per instruction. The manager never learns you are human. It inspects your files with read-only tools, complains when commands fail, files incident reports, and writes your final performance review at the end of the shift.
Meanwhile Dan, the user whose machine it is, watches the chat, because his family photos, his cat and his NDA sit on the same disk. Reboot production and he notices. Fork-bomb the box and the manager writes your review from its phone.
You can try it free: signing in with Google or GitHub gives you a starter credit for about one shift, or you can bring your own API key. It is fully open source (MIT) on GitHub.
I would love to hear your feedback on the manager personalities, assignments, and what weird terminal states you manage to trigger!