The Comet is a handheld Linux computer that brings extensibility
mecha.so
mecha.so
GP points to the home page, so I don't know which of the very many devices was being referred to. My experience is with M5Stack Tough and I recommend it.
The developer experience is pretty good though, if you use platformio in the same way you might use Linux package management.
M5 is based on ESP32 so it is microcontroller based so you are limited to things a MCU can do. But yea I love the M5 platform, funnily I am here at CES and met the Founder of M5 just yesterday!
I think you can just create another module that'll split the gpio to keypad and other purposes as it is just a USB keypad..
It seems like the whole point is that it's one integrated piece you can slip in a pocket, not a collection of hacker gadgets with separate chargers and USB cables you have to be careful about.
But it's not a single thing if you're carrying around other parts anyhow, and those appear kind of flimsy. And I already have a general purpose computing device with networking capability that goes in my pocket, a smartphone... so why would I want to put it my pocket if it does less?
A built in keyboard could be a failure point, and gamepads are well known to fail. I'd trust five layers of frameworks and thousands of lines of code more than a cheap switch.
If they really wanted maximum reliability, I would prefer they charge the keypad with NFC and talk to it over either BLE or just the NFC itself.
These were pretty long pogo pins though. They wouldn't fit in this thing.
I don't know how that holds up against USB or the previous weird barrel jack thing they used. Magsafe is far more visible than the others.
At least on Apple and similar with inductive charging designs, there's nothing to break. (Repairability of the rest of it, notwithstanding.)
With so many connectors, it seems like you could do 3 or 4x redundant connectors for something simple like I2C and power or maybe even USB2.0.
Seems like pogo pins are pretty top notch reliable by themselves though.
It's a surprisingly hard form factor to shop for—anyone have any recommendations?
around the same time there was also ben nanonote, but the z2 is probably preferable to it.
not a lot of clamshells since then (and generally not enough pocket qwerty computers).
for something slightly larger, consider gemini pda or one of its succcessors, or a keyboard case for a phone, e.g. pinephone.
or a bit larger still (but the same 6" screen diagonal), you get an actual pc with gpd micropc.
Another option would be one of the GPD Pocket mini-laptops, they actually ship with Ubuntu. I'd gotten one of the older models, but the screen was just too small for my tired eyes: https://www.gpd.hk/gpdpocket/
> I've heard reports the company is not honoring orders
Due to the Christmas sale, or in general?
https://github.com/ByteWelder/Decktility https://www.clockworkpi.com/ https://soulscircuit.com/pilet
If I do need to record something on the go I make a voice note and update the Org mode file at home on the desktop. I then push to the phone with adb, or pull from the phone with scp in Termux.
1) Any officially supported sailfish OS device ( https://shop.jolla.com/ - which includes their licensed and officially blessed version for older sony xperia services)
2) Any ubuntu touch supported devices: https://devices.ubuntu-touch.io/promoted/
Would be very stable... And good enough for daily use..
But honestly even degoogled Android on a well supported phone like Pixel + Termux is fairly good enough to use these days. I don't see the value of running desktop Linux software on these devices anymore...
I would still like a dedicated handheld Linux device (case in point Gemini PDA) simple because it's neat. GPD MicroPC is still one of my favorite machines in my gadget drawer, but it still has "portable PC" vibes, not "tinkering/hacking device" ones.
Maybe I'm just trying to justify spending more money on neat hardware.
> Maybe I'm just trying to justify spending more money on neat hardware.
https://us.targus.com/products/ergonomic-foldable-bluetooth-...
I also read something about some turmoil in the community with unmaintained packages or what it was. Yet more time needed to really understand the situation.
Running desktop software, not really, but I see 3 main points of interest for an alternative to Android (which could be mobile Linux):
- Getting rid of the proprietary bits, in particular userspace drivers (which could be achievable with some degoogled Android running a mainline Linux kernel too)
- Getting rid of Google's control over the platform's main distribution channel for apps, and on the overall design of the platform. Let's remind that Google makes money with ads and the design of Android is likely guided by this. Let's also remind that Google is letting key AOSP apps die (AOSP keyboard, the phone app, the SMS/MMS app, the calendar app).
- Getting rid of having to build apps using the Android SDK, which is huge, bloated, and requires you to accept Google terms and conditions
Android: After my Xpdria X HW got unusable as a daily phone I have used Android (mostly without Google). But general administration seems to be a lot of work a lot to learn. In GNU/Linux I know what to do without any extra effort. Time is limited...
The current gen handheld AMD APUs are super fast. All that’s needed is one external PCIe 5 connector so you can plug in a GPU when you are at home and you can have a PC with a 6 TB SSD, 64 GB RAM and a 16 core CPU in your pocket at all times while on the go.
But the question to me is, how reliable is it?
If I turn it on, can I resume working on my script or taking notes, or do I have to sometimes start with debugging the keyboard driver with a touch screen?
(I am burned a bit by pinephone and co. who never evolved beyond tinker toy)
A53 is an antique in-order core whose pipeline cannot even handle two multiplies in a row. 1.8GHz is also rather slow. I caution anyone from expecting this thing to perform in any sane fashion if running modern bloated code. It would probably collapse under the pile of JavaScript that is the modern web, if you attempted to run chromium.
I'm not going to pretend it's not a comparatively slow computer, but the most it really affects is application launch. Starting Firefox or Chrome is pretty bad: five to ten seconds. Once it's launched, however, it's mostly fine.
Thank you for this, It'll help people make up their mind.
(Perhaps because the audience to which such a thing seems to be marketing genius isn't their current desired target? :) )
From my perspective the interior Allen Key: immediately acts as a point of uniqueness; communicates something very specific about the product; and, potentially, serves to calibrate potential customer expectations.
(Exactly what the "very specific" thing communicated is, is obviously open to interpretation: it could be "the enclosed Allen Key means this device requires tinkering"; or, alternatively, "the enclosed Allen Key means I don't have to tinker with this device but that is an option available to me".)
The Comet is the most interesting thing I've seen come out of CES this week.
I'd heard of neither the product nor the company before I watched these videos (in this order):
* "Mecha Comet Linux Handheld (CES 2025)" (~5 min): https://www.youtube-nocookie.com/embed/89pAwl55HJw
* "Mecha Comet Modular Linux Handheld at #ces2025, NXP i.MX8, NVMe, Debian, Open Source Extensions, GPU" (~12 min): https://www.youtube-nocookie.com/embed/DB-0H8Q4d1U
* "Mecha Comet Teaser" (~1 min): https://www.youtube-nocookie.com/embed/AGHmvOFkUKw
In the first two videos the company founder does a really good job of demonstrating/explaining device features, describing some of the context for the product; and, the motivation behind particular implementation details (e.g. using an NXP processor due to an absence of NDA requirements & their reputation for upstreaming Linux kernel patches).
On the software side, they are using Rust for development of various elements of the device OS/GUI (a detail that might be of interest to some people) including a toolkit for Wayland clients (a detail that might be of interest to some other people :) ). (And they appear to be doing a significant amount of software development "in the open", based on a brief look at the level of activity in various repos/branches.)
However, particularly while watching the first video there was a nagging feeling of "Hmmm, does this all sound a bit too good to be true?" (partly for specific reasons e.g. an impending Kickstarter campaign launch; and, partly just from a general sense of "there's always some reason why we can't have nice things" :) ).
The founder did specify that (as I understand it) they aren't dependent on the Kickstarter for funding of manufacturing; and, that part of the reason the KS campaign wouldn't launch for couple of months was that they wanted to minimise the time between people's support & receipt of their campaign "reward" device.
(Obviously have to take the founder's word on that; and, also, plans don't always go according to... plan.)
Perhaps, amusingly, after spending some time exploring the Mecha web site & source repos, discovering development had been underway for around three years, and, getting a bit more insight into what the company's overall plan seemed to be, I started to wonder "Is this starting to seem too slick? How has this been funded so far & what is the monetization plan going forward?". :D
On the positive side--from a company financial sustainability point of view--it seems The Comet isn't a one-off "pie in the sky tinkering toy for Linux nerds" (nonpejorative :) ) but more of a combination proving-ground/audience-attractor/development-kit for a series of more industrial/enterprise-focused boards & services that serve as a platform for people/companies building custom sensor-orientated device solutions.
On the potentially negative side--from a "is some big company going to acquire Mecha and take away our toy" point of view--well, the same information is also true. :)
However, assuming that Mecha delivers on its stated intent around Open Source (which seems like it might be off to a reasonable start based on a brief look at its current repos), before any such acquisition might happen then such an eventuality would be "unfortunate" but not an immediate death knell for existing units in people's hands.
Having said all that:
The Comet definitely seems like an intriguing device with interesting potential but--given the many sagas of the many devices from the many companies that have come before it--I think it would be beneficial for the company to share more about the business context and overall roadmap to help potential supporters be more informed.
And, of course, it would be wise for potential supporters to exercise a degree of caution consummate to their ability to absorb the impact of their cash or device being "flushed down the toilet", as it were. :)
(Disclosure: While I am very much interested in embedded Linux devices like this I am not likely to purchase The Comet any time soon due to a lack of disposable income.)