Up and running
in one coffee.

Install the plugin

Omarchy clones it disabled so you can read it first. Enabling adds a robot button to the bar; Super+Alt+A opens the dashboard.

omarchy plugin add https://github.com/OmarchyFans/omarchy-fans-agent-launcher
omarchy plugin enable fans.omarchy.agent-launcher

Sign in from the terminal

A device-code flow: the CLI prints a short code, you approve it in the browser at omarchy.fans/device, and the CLI receives a token bound to your org. Your GitHub token never becomes the cloud credential.

$ omarchy-agent-launcher cloud login
  Open https://omarchy.fans/device and enter:  FANS-7Q2K
  waiting… ✓ signed in as modpunk (org: modpunk, tier: Fan)

Create a cloud agent

Or use the New agent form and pick cloud as the runtime. Same options either way: agent kind, model and sign-in, skills or a preset, the job.

omarchy-agent-launcher create researcher --runtime cloud \
    --agent hermes --preset research-analyst --size s

omarchy-agent-launcher launch researcher            # wake + attach the console
omarchy-agent-launcher launch researcher --job "…"   # unattended run, then back to sleep
omarchy-agent-launcher cloud status                  # every cloud agent, awake hours, storage
omarchy-agent-launcher cloud sleep researcher
omarchy-agent-launcher destroy researcher            # removes the machine, stops the meter

Give it secrets

API keys for the agent's model provider are stored encrypted on our side and written into the machine at boot. They are never uploaded from your laptop as files.

omarchy-agent-launcher cloud secrets set researcher OPENROUTER_API_KEY   # prompts, never echoes

Set up a Digital Twin

The twin plugin's --cloud flag uses a machine, tunnel and checkpoints from Omarchy.Fans Cloud, with nothing else to sign up for.

omarchy plugin add https://github.com/OmarchyFans/omarchy-fans-digital-twin
omarchy plugin enable fans.omarchy.digital-twin
omarchy-digital-twin setup --cloud            # picks a size, links this laptop
omarchy-digital-twin status                    # who holds the lease
omarchy-digital-twin checkpoint "before the big refactor"

Open your twin in a browser

In the Cloud dashboard open twins, pick your laptop's twin and press start under "Open in your browser", then open the terminal. It stops by itself after 30 idle minutes. Hand the lease to the cloud first if you want changes made in the browser to win.

Start a sandbox

In the dashboard open sandboxes (g x), press New sandbox, choose a fresh Omarchy or a copy of your twin, and open its terminal. Take a checkpoint before anything brave; restore it when you like.

Use a private GPU machine as a backend

Start a GPU machine in the dashboard (or with the CLI), then register its endpoint and key as a launcher backend. Every agent can then choose it, and Rix can delegate to it.

omarchy-agent-launcher cloud gpu create qwen32 --machine performance-48 \
    --model Qwen/Qwen3-32B --idle 10m --max 4h
omarchy-agent-launcher backends add qwen32 --kind endpoint \
    --base-url https://api.omarchy.fans/v1/gpu/gpu_…/openai/v1 --key ofg_…
omarchy-agent-launcher cloud gpu stop qwen32   # or let the idle timeout do it