An AI proposes.A person decides.
Three products built on that rule. Shared rooms for cross-company work, automation for teams, and a mobile IDE.
What we build
One company, three products. They share a way of working rather than a codebase.
The work an AI does has to stay legible to the person who carries the consequences of it.
That is the whole of it. A room shows a team what an agent actually did and makes someone sign for it. An automation is worth having only when the person who owns the process can still see inside it. A commit from a phone still has to land in the same history everything else lands in.
The same three moves, everywhere
Every product runs the same loop. It is the reason these are one company and not three.
The agent does the work and puts up a change. It never writes straight into the record, so nothing arrives without being offered first.
A named person with the authority to say yes says yes, or does not. This is the step other tools skip, and it is the one that decides whether anyone trusts the output.
What was proposed, who approved it and when all stay visible afterwards. An answer nobody can audit is an answer nobody can defend.
Tell us what you are building
Client work, a room for your team, or an early look at GitNomad. All three start the same way.