What a cloud AI agent does while you sleep
A cloud AI agent is a program that takes a goal in one sentence and works on a remote machine until it has reached it: it reads files, writes, runs the tests, reads the errors and tries again. It runs on a server rather than your laptop, so it carries on after you close the tab.
The loop, in three moves
An agent reads the repository to find its bearings, writes a change, then runs something that tells it whether it was right: a test suite, a build, a command. It reads the result, fixes, tries again. That loop is what separates an agent from a text generator: it gets an answer from the world between two attempts.
A remote machine enters the picture because that loop wants things you do not want on your laptop: installing packages, filling a disk, keeping a server up for forty minutes.
What changes when the machine is not yours
The work outlives the tab. You start a task, you go to lunch, the process carries on. On a laptop, closing the lid ends it, which is the real reason people do not start long tasks.
Number stops being a problem. An agent occupies a machine, not your screen. Five agents do not want five times the attention, they want a way of watching them.
And the environment repeats. The machine starts from a known image on every project, so "works on my machine" stops being a sentence anybody says.
The problem that arrives with the second agent
Two agents in one directory write the same file thirty seconds apart, and the second erases the first with nothing to show for it. The usual escape is to give each one a copy of the repository, which moves the problem to the merge.
Murmell takes the other road. One shared directory, and a claim before the write: an agent announces the paths it is about to touch, the board hands them over, and refuses everybody else until it lets go. The mechanism in full.
The agents Murmell launches
Claude Code, Codex, Kimi and OpenCode, each in a window on the canvas with a real terminal, signed in with your own account or key. No tokens are resold: the bill covers the machine and the room where people watch. Claude Code in a browser tab shows what that looks like on screen.
Straight answers
How is a cloud AI agent different from a coding assistant?
The terminal. An assistant proposes text you accept line by line. An agent executes: it creates files, installs packages, runs the test suite and reads what it said. The first speeds up typing, the second takes a task off your list.
What does an agent need before it can start on a project?
A repository, an account or key at a model provider, and a machine to run on. The rest is context: what the project does, how it starts, what must not break. An instructions file at the root of the repository covers most of it.
What happens when an agent gets it wrong?
The work is code in a repository: you read the diff and throw it away. The real risk is not a bad line, it is one agent overwriting another with nobody watching. A file claim taken before the write, enforced by the server, removes that case.