JetKVM Mini(jetkvm.com) |
JetKVM Mini(jetkvm.com) |
Not high frequency though, tops out at 45Hz
It's paranoia, but until they release the source, it's not proven paranoia.
Automated it entirely by giving Claude hands/eyes/ears so it can take complete control of the 486 using a raspberrypi 5, a usb4vc HAT for PS/2 and a VGA capture device. Adding an ethernet card and a tcp/ip boot profile for FTP, Claude then can update binaries, test them, and gather logs. Now I just steer Claude and it handles interfacing with the machine for me.
Don't even need the HDMI or video the AI will prefer serial console.
https://hardwarerecs.stackexchange.com/questions/18561/kvm-d...
The only one that met my requirements was the UGREEN 8K@60Hz DisplayPort KVM switch:
https://www.amazon.ca/gp/product/B0CFFFHFJT
Several months in without any major issues.
I can't find much, and what I find is a bit bonkers.
I ended up buying a license to Orion, using USB-C→HDMI capture card with a Startech VGA→HDMI converter (which requires USB-A power)... in order to achieve a semi-passable KVM (but, without the K and the M...).
I paired with a USB-A Hub and a keyboard and mouse and it kinda works but it's an incredibly janky setup...
... the fact that many servers still output VGA is baffling.
The hub is a very simple and inexpensive device that essentially just lets you change USB devices between computers. It has 4 ports for USB devices, and 4 input ports for USB hosts. There is a wired remote control that allows you to switch between the hosts.
The way you turn this into a KVM-like device is by having each host connected to the monitor/s indepentently. In my case, I have only two hosts, and each one is connected via HDMI cable to my single monitor. Each host is set up to detect when the USB hub is connected, and then send a DDC command to switch the monitor input over to itself.
The end user experience is: I press a button on the wired remote, which is mounted on my keyboard tray, and then after a short delay the corresponding computer is on my screen and everything works normally.
On linux, this is trivially accomplished with a udev rule + a script calling ddcutil.
On my work machine, a Mac, I could not find a simple solution that worked reliably on my specific Macbook Pro model that I would recommend, so I had an AI write a custom daemon for this purpose.
The major benefit of this "poor man's KVM" is that you do not need an expensive and finnicky piece of KVM hardware. You do not need to worry about the video chip supporting your setup, or firmware bugs. You buy one cheap, commodity product. The downside is the reliance on the hosts, and on monitor features, and having to set up the functionality manually. In my case, this is the sort of thing I basically just set up and then never touched again, so it was a fine trade-off. If I get new computers, or a new monitor, or I change which porta the computers are attached to, then I have to change the code.
Just hit a button the bottom and it swaps to the other input and I keep the keyboard / mouse plugged into the monitor and it goes where the input goes.
Actually 8P8C.
Awful marketing.
I spent a bunch of time trying out different KVMs and ran into gachas with most of them.
Networked KVMs let you remotely access a computer or server. It's like VNC except by physically plugging into the ports. This is useful because it allows you to have remote access to a server when its OS is offline, tweak bios settings, watch it boot, reinstall the OS, etc.
Most true servers have this functionality and more built in (ipmi, idrac, redfish, etc are terminology for the feature). These plugin ones allow you to add that functionality to any machine. You're not likely to see tons of them in an actual data center, but they are incredibly useful for systems without remote management built in.
He likes JetKVM. However, they appear to be sold out. The Mini seems great but I have yet to receive a preordered item on the advertised timeline.
That article pointed to ArkKVM which is a hardware clone of JetKVM but they have now released their own software stack as open source which has Tailscale support
* https://store.arkkvm.com/blog/detail/arkkvm-fully-open-sourc...
* https://store.arkkvm.com/blog/detail/arkkvm-ota-release-note...
I like it. It works pretty well. I don't see a reason to trust it any less than closed-hardware, closed-source solutions. I would trust a fully-open-source PIKVM more, however.
The tooling and documentation for setting up Intel AMT via open-source tools is meagre, so I had to roll my own. I only need a small nudge to package it for others to re-use, so reply to this post if you are interested.
The phsyical kvms just work and worth the $.
Isn't jetKVM software open source? I thought it had a public buildroot fork? Been a while since I checked, maybe something has changed.
It also solved the issue I had with entering a password on FDE (full disk encryption) systems. Previously I had to rely on a janky solution with booting into BusyBox so that the system can be accessed remotely and using that to enter the password. Now it's just plain FDE set up during Debian install - I reboot, use JetKVM to enter the password and that's it.
One note though - I don't use their online cloud service access thingy and connect to those using my Wireguard deployment. The hardware is nice, open source is nice, but I wouldn't trust that level of access to any cloud solution by any vendor.
https://docs.espressif.com/projects/esp-dev-kits/en/latest/e...
I learnt this first myself from HN, a few months ago.
(I agree that they should spell it out on the website.)
I know some ipkvm can interface with the ATX header, but that's usually an extra addon. It also requires have some access to the ATC header during installation.
I have a tiny "NAS" where the mobo is the size of credit card and very difficult to access. The power supply is external. The USB doesn't stay powered when turned off.
I cobbled together a bare NanoKVM running debian, a relay normally closed to interrupt mains power to the target, and a tiny little power supply into a 3D printed case. I have also added a diode so that the ipkvm can also be powered by the target USB for flexibility. Note that the sg200x boots with its pin pulled high with a ~50kΩ resistor (pullup). So you need to pull it down with a 1kΩ resistor. This also renders the system more robust to transient voltages anyways.
https://wiki.sipeed.com/hardware/en/kvm/NanoKVM/introduction...
I use it to switch between my work laptop and my personal PC, and for this situation 4K is a must have. That, and the choice of HDMI instead of DisplayPort, and things that would keep me from getting a JetKVM.
[1]: https://www.aten.com/gb/en/products/kvm/desktop-kvm-switches...
https://nascompares.com/2026/06/26/gl-inet-comet-x-kvm-early...
Of course putting it on 3 NAS systems (which I can even SSH into) was definitely overkill.
But as Schlock Mercenary's Seventy Maxims of Maximally Effective Mercenaries says, «There is no "overkill." There is only "open fire" and "reload."»
It started as a firmware for myself: I wanted a KVM, but not a Linux system inside one. A colleague talked me into publishing it and it grew from there. It runs on ESP32-P4 dev boards, no custom hardware.
I am not saying it is better than JetKVM. What I find more interesting is that the Mini landed on the same chip. Their stack is completely different from the Linux JetKVM, so it must be a rewrite from scratch - a sign the idea of a KVM without an OS holds up.
Have you considered using usb-hdmi capture dongle instead of using hdmi-csi. That is my idea, you bring your own capture dongle, plug it in and good to go. The esp32-s3 (and P4 i guess) support usb host, and there is uvc example from esp-idf.
TY for link.
(Not affiliated in any way, I paid for it fully!)
Personally, I also specifically do not use the vPro NIC on my Intel machine.
The harder part may be networking: it would be harder to ask them to wire it up to the network; but with Wi-Fi, I may be able to pre-fill their WPA key.
I looked at making a solution myself, but for some reason, simulating a USB client device in software is extremely annoying, and requires specialized hardware that is also very expensive. So at that point, why not just get a KVM...
I frequently switch between 1080p and 4K for gaming since I don't like resampling artifacts. Never found a KVM that can handle 4K properly even for productivity. The mistake can be as basic as assuming a 100% scale factor.
Which solutions perform well?
It used to be a time that any proper geek/hacker would obviously know what a KVM is. More the normal KVM than the IP-KVM that could be considered a marvel of our time in the end.
Because it was so common to have just a single screen and keyboard but multiple tower desktop computers. Especially at school and home. Nowadays we mostly all use laptops so it is well irrelevant and I guess only a few young reader ever saw one in the field.
The laserprinter was a terrible purchase, since new toners aren't much cloned yet, and superexpensive.
0. https://docs.espressif.com/projects/esp-dev-kits/en/latest/e...
The big “tail” on the original gives you something to build an enclosure around, and the ports are all on the face opposite the display. That shape makes it easy to 3d print a mount for a 10” rack.
This one has stuff on four out of six surfaces (display, button, card slot, ports), and the two free surfaces are at a right angle to each other instead of being on opposite sides, so this gives you little to no leeway for mounting semi-permanently without some adhesive.
https://github.com/Retrotink/PockeTerm-II
Very simple serial -> VGA device. Unfortunately no longer in production but the github project contains everything one would need ...
I've fixed that with secure boot (my own keys, not Microsoft's) and TPM2. From that host I can run a program named Tang, which operates in conjunction with another program named Clevis.
On hosts without secure boot such as RPis and other ARM SBCs, Clevis will contact Tang at boot time and decrypt the disk.
Incidentally my x86 with secure boot has intel vPro so the KVM is not needed. The remaining sisters have a piece of lost technology used for decades to solve the KVM problem: serial (UART specifically) connections.
Some seasoned sysadmins might have memories of dialing their Sun servers serial ports for remote administration.
TPM based unlock does at least still fulfill the goal of ensuring data stored to disk is encrypted so that it can't easily be recovered from a discarded drive.
At least the new ones dropped the splitter and put 2 separate ports it seems.
And I usually refer to the "IP KVMs" as a "lights-out" module/engine as I was familiar with HP's "ilo (integrated lights out) [0]. More on lights out management on Wikipedia: https://en.wikipedia.org/wiki/Out-of-band_management
I was thinking of making my own with an ESP32 + serial, as that's enough in theory, though I am unsure how well-supported serial ports are in BIOS and bootloader. Also, IIRC RS232 is ±12V, so the ESP would need a level shifter.
TFA's device looks cheap enough (well, depending on shipping to EU), I may just get one. The screen is a bit superfluous though, and I think that the ATX power button support is a must.
Pretty much any off-the-shelf SBC can present itself as a USB HID device like a keyboard. You can buy a very capable rp2040-based board for <$5.
What part of this equation requires expensive hardware?
For high refresh 4K you also need HDMI 2.1 or DisplayPort with DSC support, and a lot of "4K" KVMs don't have it.
Also, 1080p divides evenly into 4K, so turn on integer scaling in the Nvidia or AMD settings and 1080p will look sharp instead of blurry.
An annoying quirk is that I have to use the Nvidia control panel specifically to get Windows to send a 1080p signal as otherwise, it will just upscale 1080p to a 4K signal. It makes no fucking sense to me why that would be the default but whatever.
There is also the Comet 5G, which gives you cellular backup. I use it to control my OpenBSD home router when traveling if my Internet connection goes down.
Also they say they are open sourcing the reference implementation of the "JetKVM OS Services" and the firmware of the "JetKVM Mini".
My point is more that it is way easier to make mistakes in source than it is to review said source in a project like this. So, to me, it is kind of risky to use something very green that is meant to be used as a network appliance more or less.
I expected that kind of potato IP-KVMs to surface from crypto mining space, but alas, what we got was this thing which predecessor have reputation of crashing. That might not be stupid, but a bit frustrating.
I agree that it's insane it works this way though.
In that case the path of least resistance becomes something like a $22 NanoPi NEO (has onboard ethernet, widely used in PiKVM setups)
Maybe I'll look into that, wouldn't require messing around with the picoprobe. Thanks!
It is either on my LAN with that server available, or you will need to enter a key.
It am only trying to prevent casual snooping if it goes missing from my garage though. Anybody more sophisticated can have my garage YouTube browsing history.
As in, no problem, I will just buy another one. At these RAM and storage prices, maybe not.
Except it’s an initialism
It’s a YABA
ha. I guess I'm one of today's lucky 10000[1].
Note that if anti-cheat is a concern, common dev boards like Arduino Leonardo and Pi Pico could also trigger flagging, so maybe that needs to be done slightly differently too(assuming the intent is good and this don't count towards anti-cheat evading in whatever you're doing).
As for dev boards, it wouldn't ever show up as a dev board to the PC, it would simply be simulating a dumb hardware keyboard. That's one of the reasons I wanted the board itself to have its own wired network.
How? I and many other have been using it for years without problems.
> it is absolutely not lower latency than the gpu's native display output.
and subsequent encode