Fair enough, but how many of these are as polished as their Android counterparts? How many are still actively maintained and will be for the foreseeable future? Android already is "Linux for mobile devices" and has all the features people expect from a phone. Reinventing the wheel on the software side seems counterproductive when what we need is simply more hardware to run Android on. I wish Furi collaborated with Graphene or Lineage and released a proper FLOSS Android device instead of going the "desktop Linux" route.
You're partially correct. Ubuntu Touch was discontinued by Canonical (the company behind Ubuntu) in April 2017.
However, Ubuntu Touch was picked up by the UBports project (that by now has their own foundation) and has been continued to this day. Currently they are preparing a Ubuntu Touch release based on Ubuntu 24.04 (moving on from Ubuntu 20.04). See https://www.ubuntu-touch.io/ for more.
Regarding the FLX1(s): FuriLabs worked on a way to support Ubuntu Touch apps (that can be found at https://open-store.io/) natively on FuriOS. It's also possible to boot Ubuntu Touch on their FLX1 hardware.
I think seba_dos1 refers to PureOS [0] (for the Librem 5; PureOS is essentially a FSF-endorsed Debian) and Mobian [1]. Both use a close-to-mainline kernel, Mobian's goal is to "bring Debian to mobile devices. Over time, the idea is to minimize the Mobian specific pieces by “upstreaming” customization to the original projects."
I maintain a project [1] that tries to collect and track them all, and ... at least for keeping track of how they all develop, there's not a few apps. ;-)
You currently list 720 apps, Google Play Store has 3.3 million.
Even if we assume that a good chunk of that may be "duplicate" in terms of functionality (e.g. todo apps), that is still just a completely different dimension of apps and use cases covered by android natively.
That said, Google Play is not really the thing to compare this too. F-Droid could be. Summing up the "Show all ..." counts, F-Droid clocks in at 6147 apps, and it started way earlier (2009? 2010?).
Some of these F-Droid apps (specifically those created in QtQuick or Flutter) should also run with very minor tweaks on #MobileLinux.
Also, there's more than these 720, some are just very hard to evaluate (because they are for hardware or services I don't have/use), which then keeps me from adding these apps.
I don't recommend using the vast majority of mobile apps, which exist for data-mining and addiction-maintenance purposes. Android is specifically known as the worst ecosystem in this regard. Browsers continue to work, despite the corporate push to dark patterns.
It's Linux, if you need something you or someone else will eventually write it. But first there needs to be acceptable, working hardware. Enter the FLX1/s, and we come full circle.
Sure, the majority of mobile apps suck. But even then, 1% of it is still orders of magnitude larger than the number of "Linux" apps, and there are surely more actually decent mobile apps than that.
Besides, open-source doesn't have to be GNU, there are a bunch of open source apps on F-Droid.
I also reject non-free apps whose main purpose is to snoop on their users like Facebook. But what about dozens of FOSS projects like Home Assistant, Jellyfin, Nextcloud, KeePass and many others? All of these have polished clients that work perfectly on Android and provide an experience that rivals or surpasses commercial offerings. They took years to develop and I don't expect Linux phone versions will ever be released because of the chicken and egg problem. You can emulate these apps with Waydroid on a Linux device, but you will almost certainly lose some features like GPS access or notifications. At the end of the day, you will mainly be running Android apps anyway, free or non-free, so you are better off with an Android device in the first place.
I currently use 0 Android apps, so honestly believe one could survive without them.
I'm glad Waydroid is a thing for folks that are, and sorry attestation prevents full compatibility. But neither issue should get in the way of investment in modern Linux phones, which I've been desperately waiting for, for years.
Web apps are still a thing as well, believe it or not.
I think that's because Sebastian develops software for the Librem 5 and thus is likely not running a stock PureOS Byzantium.
I managed to get similar results in my holidays in late August(on postmarketOS 25.06 with Millipixels 0.23.0 and a patched kernel [1]).
Please note that I just used the Librem 5 as my "mainline Linux camera" and have not had a SIM in it in a while, and thus can't comment on how well postmarketOS 25.06 would work for daily driving.
The only major thing these kernel changes introduce is RAW10 support, which helps photos with high dynamic range that would have visible banding in the shadows otherwise. There are more changes, but their effects are either subtle or unrelated to still photos. Most of the pictures I post have been taken without these changes, only some of the most recent ones used them.
When installing just two apps, even if both are in the same (KDE or GNOME) realm, you can very easily end up with 8 flatpaks (including runtimes) or more. This is due to a variety runtimes and their versions: One for KDE or GNOME Platform release (about two a year) plus a yearly freedesktop base) and not all apps being updated to the latest runtimes constantly.
You then have to add at least 6 locale flatpaks to these hypothetical 8 flatpaks.
Especially with Debian, locales matter, of you don't do a `sudo dpkg-reconfigure locales` and pick what you need before installing flatpaks on a default install, you will get everything and thus gigabytes of translations you don't even understand, wasting your disk space.
I follow the ecosystem for two projects [0], and the app ecosystem keeps getting better. Also, there's always Waydroid (LineageOS equivalent of Android 13 in a Container) and Android Translation Layer [1].
The best way to experience #LinuxMobile is IMHO on former Android Devices like the OnePlus 6 or Pixel 3a, as these are well designed devices with SoCs intended for phones. While having native hardware would be great, it's really hard to produce relatively bug-free hardware from scratch. With mainline, the SoC is always going to be older, and small-scale production leads to high prices. The economics just don't work out for non-enthusiasts who then ideally are also up for paying a software support subscription. Also, as a manufacturer you'll want to somewhat control the experience, leading to 'yet another small scale device-specific distribution, which has it's downsides.
Across devices audio issues are unfortunately still somewhat common, in part caused by the Pipewire transition (things that had been figured out for PulseAudio need to be figured out again). But it's on people's radar, and funds are invested to solve this [2].
VoLTE is another challenge, as it allows to carriers to become more gatekeepy with their stupid device support lists - of course there are workarounds like using a SIP account, but that's too inconvenient for many. Also, RCS support is not yet there, but as long as SMS/MMS fallbacks are still around, this not a dealbreaker.
"No one was working on it for year" is both untrue and unfair. There have been continued improvements (both PinePhone and PinePhone Pro), and I know that there are people who have put in the midnight oil to make them happen.
What that work was not is 'paid for by PINE64'. It's also not been enough to raise the bar enough to make the phone work well enough; but if you consider what's involved there, it makes sense.
You don't just need to write/fix a driver, you need upstream (or at least a distribution) to accept it and include it for that work to make a meaningful difference for anyone else.
It did not get far in market share, sure, but it never stopped going like Big Screen did and made the transition to 6 at the same time as Plasma Desktop.
Here are some recent developments:
1. https://plasma-mobile.org/2025/07/08/releases-25-07/
(also explaining that most software is part of 'Plasma Gear', which makes sense as it is software for both mobile and desktop, but has the downside of making Mobile less visible)
2. https://blogs.kde.org/2025/07/05/this-week-in-plasma-chuggin... and look for Plasma Keyboard in this post.
(Personally I am a bit sad that they are seemingly abandoning Maliit, which is also used in Sailfish OS and Ubuntu Touch and is IMHO fine, but hey, at least it shows that Mobile is not dead.)
I mean more like it didn't progress beyond prototype designs. It never felt polished or even close to nice looking. More like desktop Plasma UI minimized to touchscreen scenario, rather than significantly designed for it (like Sailfish). But good to see it's progressing too.
Yet, there are a select few non-gui flatpaks.
One I use frequently is Zola [0], a static site generator that I use because it's significantly faster in building my site with 'zola serve' than the native Alpine Linux package. All I had to do to make it convenient was aliasing flatpak run org.getzola.zola to zola.
Beg to differ: https://linuxphoneapps.org/
(And no, that does not list all of them. Only all I got around to adding.)