Skip to Content

Get started

CMDOP runs an autonomous AI agent on every machine you own and gives you one place — a desktop client, a CLI, or your IDE through MCP — to talk to all of them. This section is the on-ramp.

A suggested reading order

If you are brand new, walk this list top to bottom; each step is short.

  1. What is CMDOP — the one-page pitch.
  2. How it works — the moving parts in plain language (daemon, relay, workspaces, machines).
  3. Architecture — the block diagram for visual learners.
  4. Install — pick a surface (Desktop, CLI, headless agent, mobile).
  5. Quickstart — first chat in under five minutes.
  6. First connection — register a remote machine like vps-audi, attach to it, run a command.
  7. First automation — one trigger or skill that does real work.

Pick your install path

GUI on macOS, Windows, Linux. Recommended for day-to-day work — seven tabs, global hotkey, per-machine inspector chat.

One binary, scriptable, runs anywhere. Same agent core as Desktop. cmdop chat, cmdop connect, cmdop run.

VPSes and servers. cmdop register once, then cmdop agent start keeps the machine online.

iOS and Android — a companion client for monitoring and quick prompts on the go.

Background reading

OAuth, the permission gate, the floor, and where tokens live per surface.

SSH, Tailscale, Ansible, and Claude Desktop alone — when each makes sense.

Concepts you will hit early

  • Workspaces — billing, membership, and credential scope.
  • Machine identity — how CMDOP resolves vps-audi to a stable UUID.
  • The daemon — the long-running process that keeps a machine “online”.
  • Permissions — what the agent may do without asking.

Where to go next

Last updated on
What is CMDOP