Rooster Wake

Rooster Wake is for a computer that should be asleep by default, awake while it is used, and reachable when you need it — answered by name from any Claude chat.

With one laptop and nothing else on the network, Claude can put it to sleep from anywhere, hold it awake while something long runs, say whether it is up, and, on Windows, start Claude Code in a folder you registered in advance, handing you the live session in the Claude app. The agent keeps it up while a session, a transfer or a run is actually moving, and lets it sleep again ten minutes after the last activity.

Waking a machine that has slept is the one part that needs a second device. A sleeping PC cannot wake itself and the agent on it is asleep with it, so the wake comes from something else on your network that stays on: an Amazon Echo you already own, through the free skill; the same free agent on a NAS, a Home Assistant box or a Pi; or an open-source dongle you build. Until one of those exists, sleep it from Claude and wake it at its keyboard or by its lid.

Ten tools: list machines, machine status, wake, set up a PC for waking, hold awake, release a hold, start a Claude Code session, stop one, sleep (which always asks you to confirm), and a one-time sign-in link to the dashboard. No shutdown tool, deliberately.

Getting set up: once connected, ask Claude to "list my machines". With nothing installed yet it answers with the one command that installs the free agent on the PC you want to reach, already filled in with your account. Run it at that PC, ask again, and the machine is there. Remote AI sessions also need Claude Code installed and signed in on that machine (your install, your Anthropic account — we start and stop the process and never see the conversation). Waking is free on every plan; the free plan includes ten sleeps and ten Remote AI sessions a month, and the paid plans are unmetered and add the hold, restart and shutdown. Full walkthrough: https://roosterwake.com/run-claude-code-remotely