Hello,
I’m looking for someone to build me a totally custom keyboard, chassis and all. I have the layout I want and it has 122 keys, not including the media keys/volume roller and macro selector. All the sites I see online so far are for standard PC layouts or variants of them. I want a true custom keyboard of my own layout and design, so if anyone has any suggestions, please let me know.
US based by the way.
Thanks!
So, you’ve got a non-trivial ask here. It’s certainly doable, but it’s going to start in the hundreds and go up very quickly from there. Feel free to give us your budget and requirements for components, materials, and build sophistication, but you might get more nibbles with a “WTB” post at a place like Reddit’s “mechmarket” community.
I admit, I am curious about the custom 122+ key layout though. You are right at the edge of what can be done without things getting unpleasantly interesting. 
And welcome to keebtalk!
I have a mock-up from keyboard-layout-editor and I realize I went a little overboard. Obviously, I can bend a little, I was just wanting to start at the dream and whittle away to realistic when I found someone that can craft it.
What I’m looking for is a chassis that is similar to the Logitech G910, but instead of the Arx dock tray, just a cellphone slot like you find on some keyboards. I want it to be a mechanical, again like the G910, and something that will last.
Is there a way to send you the link to the layout?
You’re all good, LOL. Why are we here if not to find just the thing we want? 
You can send the link using the “permalink” button in KLE, or paste in the “Raw Data” tab into a blockquote, but most common is to just throw up a screenshot or the image from the Download button.
Everything will come down to when you need it, what materials you need it made from (e.g. CNC aluminum case is an order of magnitude more expensive than 3D printed), how complex the design itself will be, and the methodology for construction (bespoke circuit-board is more expensive than hand-wiring). Keycaps, stabilizers, and switches will be the other big expense to consider.
I’ve hand-wired two boards roughly as large as your design, so I know it can be done. However, one was simply re-wiring an existing vintage “battlecruiser.” The other is made from laser-cut painted hardboard and 3D-printed plastic, and its large but compressed layout was already taxing my setup. There are some extremely talented makers around here, to say nothing of the broader community, where there are entire discord servers dedicated to designing boards, and people who’ve run multiple successful group buys can often be found just hanging around.
There’s frankly not a ton of money in it for most of them, hence my suggestion to post a “Want to Buy” listing somewhere with broad reach. If your pockets are deep enough, you can almost certainly get exactly what you want.
So the design I want is the v4, but I want the chassis to look similar to the G910 but with a phone slot like the one attached as well, so let me know what you think. Obviously, I want mechanical switches for the main keyboard, but can work with whatever we need to keep the cost down a bit. Budget wise, I am willing to spend some money to get what I want, but if it’s a crazy dream, it’s a crazy dream lol.
So that is indeed a big fella. It’s not impossible, though, not by a long shot. If you can live without addressable lighting (i.e. none at all, or you just turn it on), it looks like you need up to 28 GPIO pins to code it using the standard toolsets and “usual suspect” hardware, and fortunately, that’s about where they top out. There are other tricks like Japanese matrix wiring and shift registers that could reduce that quite a bit, but the usual method is to wire buttons into a matrix that has diodes along one axis to prevent ghosting, and the max number is the square root of total keys, times two, rounded up if it’s not an integer. Then you’d need a pin for each lock indicator, and 2-3 for a knob.
A printed circuit board of the cheapest type would require a spot for a large format MCU, 140ish diodes, and a few resistors. Once designed, it would be about $60-$200 to have the minimum order of 5 of them made by the cheapest vendor and shipped to the US (thanks, unpredictable but likely huge tariffs!). You could have much of it preassembled for more, but I’ve never done that myself, or even priced it out; I don’t think it would go past double for a keyboard PCB though. Handwiring is potentially less reliable, but it’s robust in the sense that it’s easy to fix. It can be done for the cost of tools components, and wire; occasionally (though not with MY boards, LOL) it can verge on being a work of art.
A switchplate is nominally optional, but unless you do a very specific type of mount, I’d say it’s close to mandatory for a board this size. I haven’t priced these out recently, but on this size figure another $100 at least. Lots of options in this space, though. Don’t take that number as gospel. Aluminum, FR4 (circuit board material), or various plastics could all work. The cost is more in the setup and machine time than the material.
That leaves you with a case. a CNC’d case for a board this size will be extremely expensive, having to be a bespoke design and quite large. My initial guess is that prototyping one would be around $500 at an absolute minimum, and maybe a good bit more these days, though I’d be extremely happy to be told I was wrong. 3D printing could also be an option, and is probably the only other thing that could get you close to the aesthetic sense of the Logitech, but you’re looking at either the “traditional” 3D printed surface, some creativity to design for the newer surfacing techniques off the machines, or a lot of post-processing, which can get pricey if someone is doing it for you.
Stacked laser-cut acrylic is potentially a more attractive option, and also opens up the number of vendors/makers who can help you. You could also use the same vendors you selected for your plate and get something that could be stacked or bent and maybe have some sort of cushioned mount. Neither of these will LOOK like the Logitech though. There’s also the super simple “sandwich” idea, which is barely a case at all, but rather some structural means of separating the switch-plate and a base plate; these typically leave the sides of the switches exposed, which is still fairly common in gaming boards, but less so in “enthusiast” boards.
Your layout is mostly fine, though no “keyboard” switches will really fit under the smaller buttons you have, and the trend in the hobby is to simply make everything a normal sized key. I assume you’d like volume in a rotary encoder, and most of the more common ones will also serve as a button that can be used as mute (or anything you like, but that’s common). You could run into some issues finding keycaps with the right legends for your oversized escape keys, but you were smart to pick 2U, since at the very least the right size and shape for that placement should exist.
Spitballing again, but I’d say the absolute bare minimum BOM for a hand-wired proof of concept for this board, looking a bit like a flat version of older Hyper7 boards, and using commodity caps and switches, and completely excluding anyone’s time, might be around $300, looking something like this:
You’d probably add $100+ (as discussed) to change to a simple soldered PCB that piggybacks the MCU, or some amount more for a properly designed one with integrated MCU, and more again if you need hot-swap, though you could erode the $40 for other supplies somewhat. You could enclose the sides with something laser-cut or 3D printed for maybe another $30-$50 if you design it with an eye towards inexpensive fabrication. This would NOT look like a G910, but it would bring your layout into the real world. The look can be nce enough, but it’s very specific to what it is. “Pretty” gets expensive really fast.
Would you be interested in fabricating it?
Thank you, but while I enjoy the noodling, I’ll say no. First, remember the part about time. I value mine at zero for my own builds, because of the opportunity to learn and proceed at my own pace and the minimal stress of bearing the risk for the outcome, but it would be expensive for someone else’s, and potentially in a way that wasn’t fair to them as a “customer.”
Because… I am still very much on my journey as a maker, and while people on Keebtalk in particular are very understanding that we’re all engaging with our hobby in our own way, my builds are idiosyncratic to say the least. For instance, I completely forgot to mention stabilizers (ten bucks for cheap ones, btw) because I often do builds that skip them. 
I do genuinely hope you find someone who can help, or that you take it on yourself to learn at least some of the design parts, and that this discussion helped with perspective on what can be done, how, and at what cost. I will occasionally see people who want to get something truly unique and they’re simply slapped down with no explanation, and that’s not fair to them. Yes, this is expensive. Yes, getting help makes it even more expensive, but it’s possible, and if you go into it armed with the right expectations, you can make an informed decision.
Feel free to keep posting, and as I said, don’t be shy to reach out to the broader keyboard community on Reddit or Discord.
insert round of applause/standing ovation for the very thought out replies
