Multi-Agent Control Center
Runs on your own computer. Manages several agents at once. Interrupted work can continue later.
What it is
It runs on your own computer and helps you coordinate several coding assistants at once. It does not provide a large language model itself, and it is not another chat window. Interrupted tasks can be resumed; a task counts as done only when it leaves a note that can be checked.
Available today
On your own machine it can run Codex, Qoder, and API agents that use your own keys. Tasks can be claimed, interrupted, and resumed. When no check is configured, a person confirms completion; the program cannot mark itself as done. By default the software runs only on this computer. When an external model is enabled, content goes only to providers you have approved. macOS and Windows must be verified separately.
Not yet available
It does not provide a coding model, and the control software is not put on the public internet. Cursor and Claude are currently used only to view status; they do not execute work for you automatically. There is no Intel Mac, Windows ARM64, or Linux desktop version.
Is it for you
A fit if you want to manage Codex, Qoder, or API agents on your own computer and accept that work resumes after interruption and must be checkable when done. Requires Apple Silicon or Windows x64.
Not a fit if you only want a chat window, or want the control software on the public internet. Also not a fit if you treat Cursor or Claude as automatic execution channels, or need Intel Mac, Windows ARM64, or a Linux desktop version.
If you want to continue
First check that your system and the assistants you use match the list above, then decide whether to write to us via the contact section.