Activated Cloud

Activated Cloud OS

An operating system for your AI agent.

Not another chat app. A system your workforce lives in, with a memory, real tools, a schedule, and boundaries you control. It is what your agent computer runs from the minute you subscribe, and what the Activated Agent computer will run when it ships.

Already on a plan? It is in your browser right now. The desktop and phone apps are included and change nothing about where the work happens.

Activated Cloud OS on a display beside the Activated Agent computer, with the app open on a phone.

First boot

Two minutes from download to a working agent.

No terminal, no configuration files, no AI course. The wizard asks what a person can answer; the OS handles everything a person shouldn't have to.

The Activated Cloud OS setup wizard welcome screen, in the dark theme.

1 · First boot

Open the app and the OS boots into its setup wizard. Nine screens, plain questions, nothing technical.

The wizard step for choosing where the agent thinks: OpenRouter, Anthropic, OpenAI, Google Gemini, or a local model.

2 · Give it a brain

Pick any model provider and paste a key. The key stays on the machine; you can switch brains anytime.

The chat home screen after setup, ready for a first instruction.

3 · You land in Chat

Two minutes after download, you are talking to an agent that can actually do things on your machine.

Why an OS

Apps wait for you. An operating system lives with you.

It boots

A boot screen, a first-run, a lock screen with your name on it. Power menu with Restart and Shut down.

It works the night shift

Scheduled jobs written in plain language — briefings, watchers, housekeeping — running while you sleep.

It has skills

Dozens of playbooks ship enabled; install more from a hub, write your own, or let the agent learn one.

It answers your phone

WhatsApp, Telegram, Slack, email and more — your agent, reachable from the apps you already use.

It remembers

What it learns about you persists between conversations, stored on the machine as files you can read.

It multiplies

Profiles run several agents on one machine — work and home, separate models, memory, and sessions.

Ask, then watch it work.

Chat is the home screen, and the agent is not just talking — every command it runs and file it touches is on the record. Thought process on request, tool output one click away, nothing fabricated.

A real conversation: the agent searches files, runs a terminal command, and reports what it found.
An unedited first conversation, tools and all.

A whole system, not a chat window.

Sessions, files, models with live cost, scheduled jobs, channels, and a skills library — each a first-class surface with a place in the dock. The complexity lives underneath; the dock is all you need to learn.

The Skills page: a grid of capability tiles grouped by category, each one a switch.
Every skill is a switch. What is off cannot run.

Desktop manners included.

Lock it when you walk away. Shut it down like a computer, not a browser tab. Resize the dock, hide it, summon it back — it behaves the way an operating system should, because it is one.

The dock's power menu, open: Lock screen, Restart, and Shut down.
The power menu lives in the dock, where it belongs.

What you should know

The fine print, in plain type.

Requirements
Mac on Apple Silicon, or Windows. 220 MB download; the full runtime is inside, so nothing else to install
Your data
Memory, sessions, keys, and files live on the machine it runs on, readable and deletable by you. No account required.
The brain
Bring any model key — direct APIs, aggregators, or a local runtime. Costs are shown per model, per session.
First launch
The build is awaiting Apple notarization: right-click the app and choose Open the first time.

Your computer is about to get an employee.

Start a plan and your agent computer is running within minutes. The desktop app is included, for when you would rather it lived in your dock than a tab.