An IP KVM for a homelab
Nothing in a homelab has a BMC. So when a machine stops booting, the fix is a monitor, a keyboard, and a walk to wherever you put it. This is that walk, removed.
The evening you lose
You know the shape of it. An update to Proxmox, a kernel that will not come up, a NAS that answers pings and nothing else. SSH is gone, so the machine may as well be in another country.
What you actually need is thirty seconds of looking at a screen. What it costs is finding a monitor, finding a keyboard, unplugging something from your desk, carrying it to a cupboard, crouching on the floor, and doing it all again when you realise you also need the HDMI cable.
An IP KVM is that monitor and that keyboard, permanently attached and reachable from the sofa.
It pays for itself the first time
Not in money — in the evening you get back.
The first time a machine will not boot and you fix it from a browser instead of crawling behind a rack, the thing has already earned its price. Everything after that is free, and the device does not stop working at the end of the year.
If the machine is not in your house — a family member's, an office, a friend's rack you look after — then it is money too. One drive across town costs more in fuel and evening than the hardware does.
Work out what one trip costs you →
What it does that remote desktop cannot
Everything that matters in a homelab happens before the operating system is ready to talk to you:
- A GRUB menu waiting for a selection nobody is there to make
-
No bootable devicebecause a USB stick was left in - A filesystem check that stopped and wants the root password
- Boot order to change after a disk was added
- Virtualisation extensions to turn on for a hypervisor
- A kernel panic you need to read before you can plan anything
- A fresh install, from the installer's first screen onwards
Remote, without opening your router
Homelabs sit behind whatever the ISP gave you, which increasingly means CGNAT and no inbound connections at all. Beacon dials out and the session is relayed, so it works from a phone on mobile data with nothing forwarded, nothing port-mapped, and no VPN to keep alive.
On your own network it connects directly instead, which is quicker and keeps working when the internet does not.
The machine stays yours
Nothing is installed on it. No agent, no driver, no account, no kernel module, no service to keep updated, nothing to break the next time you rebuild it. To the machine, a monitor is plugged in and somebody is typing.
Which also means it works on the things that cannot run an agent: a firewall appliance, a switch with a serial-free BIOS, an old box running something you would rather not touch.
Handing a machine to somebody else
Sometimes the problem is not yours. Send a link, and one person gets that one machine in their browser — full control, or watch only — without an account and without access to anything else you own.
Common questions
My server board already has IPMI. Do I need this?
No. IPMI, iDRAC and iLO do the same job and they are built in. This is for
everything else: desktops, mini PCs, consumer boards, the machine that used to be your gaming PC.
Does it work with Proxmox, TrueNAS or Unraid?
It does not know what they are, which is the point. It shows the screen and
sends keystrokes, so it works at the installer, the boot menu, the shell, and the web UI's login page alike.
One per machine?
Yes — it physically plugs into a machine's HDMI and USB. For several
machines, a cheap HDMI and USB KVM switch in front of one Beacon covers them all, and hotkey switching works through the emulated keyboard.
Can I try it before buying anything?
If you have a Raspberry Pi 4B, the software is free and does everything the
hardware does — see the comparison.
What about power? Can it switch a machine on?
Not on its own — it replaces the monitor and the keyboard, not the power
button. If the machine is off, a smart plug and the BIOS setting that restores power after a loss will do it.