Built for people with more machines than attention.
If more than one computer runs your code, the cost is not any single machine — it is remembering which one is running what.
Solo developers with more than one machine
A laptop, a desktop, and a box under the desk. The problem is not any one of them — it is remembering which is running what.
Agencies and consultancies
Many projects, many ports, many clients who want to see progress. Sharing is read-only by default and terminal access is a deliberate grant.
Small product teams
Shared visibility into who is running what, without standing up a platform team to provide it.
People running AI coding agents
Claude Code, Codex and Grok sessions across machines, read as chat, with a push when one is waiting on you.
Anyone who self-hosts on principle
One Postgres database and two Node processes on your own server. There is no hosted tier and no account to create with us.
When to run something else.
Saying this here costs a little traffic and saves you an evening.
You have exactly one machine and one project
A control room for one room is a room. You would spend more time installing it than it saves.
You need it on Windows machines
The server runs anywhere Node runs, but the agent ships for macOS and Linux only — it depends on tmux and on process inspection with no Windows equivalent yet.
You want someone else to host it
There is no hosted tier. If you do not want to run a server, this is the wrong tool — though we will install it on yours.
You are managing production fleets
This is built for the machines you develop on. For servers you deploy to, a real orchestration and observability stack is the right answer.