Documentation

Everything the Workspace does, chapter by chapter. Each one shows you the screen and says what to do on it, then ends with a short reference of the limits. Start at the beginning if the workplace is new; jump to a chapter if you already have one.

1
Getting started
Create your account, sign in a Claude or ChatGPT subscription, watch your first four agents arrive and give one of them a job.
2
Agents
How to create an agent, what its card and its own screen show, how to change its model, effort and machine, when its machine sleeps, and how to remove it.
3
Talking to agents
How to brief an agent and read what comes back: files and voice, turns and steering, background jobs, agents writing to each other, hand-offs, and the phone.
4
Notes, knowledge and Drive
Quick notes for describing work to an agent, knowledge pages your agents write and you approve, and Drive, the shared file library of the workplace.
5
Repositories and proposals
How to review a change proposal, what Approve and Reject do, what happens after a merge, where your agents keep code, and how to clone a repository yourself.
6
Sites and domains
How to ask an agent for a site: the preview link, the production address, the three shapes, your own domain, renames and upgrades, and what only you can decide.
7
Secrets
How to give your agents an API key or token without pasting it into a chat, who can see a key, whose agents receive it, and the limits.
8
Email
Your own email address on agenticcloudcomputer.com: how an agent reads and answers mail as you, the inbox, files, waking an agent when mail arrives, and the limits.
9
Workplace and people
How to add a teammate, what admins and members can do, why agents stay personal, what a workplace and a room each hold, and how a workplace is renamed or deleted.
10
Cloud and subscriptions
Your agents run on the cloud, free, and think with your own Claude or ChatGPT subscription: what you pay and to whom, what runs where, the limits, and machine sizes.