This isn't competing with F-Droid, it IS F-Droid. Everything from the statistics to the APK downloads (notice the redirect) are sourced from https://f-droid.org/.
It's a beautiful front end and necessary improvement over what exists, but it's complementary not competitive.
1. Link to download button to the app entry in F-Droid so I can download it from the Android client. (Getting the user to trust random APKs is not a responsible behavior.)
2. Clarify that this site is a viewer for F-Droid and non-UI contributions should be sent there.
Other misc notes:
Install F-Droid as a system app (root) in order to install packages using the client without having unknown sources checked.
Material design for the client is coming 
It's awesome to see this on the front page of HN, If you think these things are worth supporting, do so!
Being open source, you're free to do whatever you like, but I suggest you talk with the F-Droid devs to make sure everything is mutually beneficial.
Always imagine someone stumbles over your page, not knowing what it's about at all.
1. HTTPS, especially for APK downloads
2. Information about the APKs: built/signed by whom
When updating, the system checks, whether the newer APK is signed by the same certificate and refuses update, if the certificate differs.
It simply does not matter, whether your cert is vetted by CA or not. So using PKI would not make sense there.
The rationale was not forcing the developers to purchase signing certificates in order to publish for the platform. It does not make difference anyway.
At least even MacOS has a App Store + Identified Developers. Though of course, iOS doesn't have that..
I don't quite understand. How do you turn off the Play store?
It is necessary to make a distinction between what you want to believe and the reality. The reality is, that requiring validated keys would put the keys to the "official Android" kingdom into CA's hands. In addition, because Android is an open-source project, any alternative distribution would disable that. It would cause real fragmentation of the platform, where apps would run on one distribution and not on another, the difference would be only the signature. Google (correctly) decided, that they do not have to fight this fight.
A side note: getting CA verified can be problem in some parts of the world. What if you are Chinese? Crimean? You can still use Android as it is; you can't use any platform, that requires to be "Identified" by CA.
Having said this, they definitely seem to be useful as a general rule of thumb about how apps are more popular, _relative to other apps_.
 - https://gitlab.com/fdroid/fdroiddata/blob/master/stats/total...
 - https://gitlab.com/fdroid/fdroiddata/blob/master/stats/total...
1. Couldn't find an option to add my app.
2. Screenshots of apps would be nice (some people might be more interested in the app itself than in the source code).
You can add your app by submitting it to F-Droid. This is just a front end.
Your site looks great and I don't mean to take anything away from it, but when I see something like this I think of it as merely an art project, something I might see on Dribbble. I would never download APKs from here for example, but have no problem trusting F-Droid's APKs.
It'd be nice to add a link the the f-droid page for each application, where such a page exists.
The relationship between LLVM and GCC is an example for this aphorism.
LLVM exists because no amount of contribution to GCC will accomplish LLVM's goals (licensing and modular design).
Natural selection. The losers die so the winners can take up more ressources. That'd be a contribution to the shared goal.
> LLVM exists because no amount of contribution to GCC will accomplish LLVM's goals (licensing and modular design).
The reason why the competition exists doesn't matter. But the competetive pressure that results from its existence does.
Imagine there is an open-source project. Its development is active, many people are contributing patches, all is well.
Then, the maintainer disappears. Forks start to appear, but as there is no centralized development any more, all you can find is a dozen forked GitHub repositories, each with a few improvements - but, without getting your hands dirty, you have to choose only one.
Eventually, someone comes along and rallies together the forks and revives the project. You may think that all is well again, but suddenly the original project maintainers return and continue development in a direction incompatible with the main fork. Neither project's leaders show interest in uniting the forks back together.
What's worse, now you have a bunch of other open-source projects that depend on the original version, and a bunch which depend on the fork. In many cases (e.g. libraries) you can't use both the original and the fork simultaneously, meaning you can't use any components that depend on both forks in the same project. And now you have a huge mess on your hands.
This or similar situations have happened many times before. I can name 3 projects that have suffered this fate off the top of my head, projects that I've been personally involved with.
Fragmentation is bad. Forking should be a last resort.
But fragmentation is going to happen and occasionally something really better comes along that wouldn't have from improving existing things. So I'm just as interested in making 'natural selection' work better. Now, it's extremely hard for it to operate, because it's hard to discern differences among similar projects, let alone which is better. Popularity data might help a little bit.
In development i can load the APK directly to the watch, but how will my users do it?
Android Studio packs the wear.apk inside of the mobile.apk but your phone won't upload the wear.apk to your watch unless it comes from google. (at least in my experience.
2. Click it twice,
3. Resize the browser to its original size,
4. The sidebar has vanished forever.
The interface took a little while to become used to, but in general I've found it to be an excellent GPS navigation tool.
Maps are downloaded, so everything works offline. I've used it extensively around Europe and especially in Austria.
I drove from Graz to Vienna last week. I know the way generally, but didn't really know where to go in Vienna. OsmAnd navigated me to where I needed to be, no problem. Then back out of the city and on to Krems. That's about a 275km journey.
I have had the app crash a few times, but that's only been when I'm out walking in the hills and using it as a walking GPS. Even though it's not open source I now use ViewRanger for this instead (apk available directly from their website, so you don't have to go via Google Play, although it is available there too).
I've used OsmAnd very successfully on both a Geeksphone Revolution and a Motorola Moto G 2nd Gen.
My only suggestions would be to make the source and tracker links more prominent and flip newest so its first or consider making separate menu points for each.
Also any link to the fossdroid src, quick google search for fossdroid didn't turn up anything for me.
The quality of Fossdroid/Fdroid apps are hit and miss - some are good, some not. I should know, some of my ports to Android are on it ( http://fossdroid.com/a/truly-creative-live-wallpaper.html ) where I'm still very nervous that critical sections survived my port safely.
Reading a list sideways feels bizarre: http://fossdroid.com/c/system/
Intuitively I feel like my eyes are trained to flick sideways to read a short line and then down. Recalibrating to flicking sideways to read a short line, then moving sideways again a variable distance feels difficult.
Will try and learn how to make my own apps by reading the source of others.
One thing that MUST be addressed before considering it something that people actually use is the total lack of HTTPS. This is not optional, and indeed should be mandatory like with https://f-droid.org and https://play.google.com.
I'd love to see more thinking about the detail view of the app. For example, there should be a "buy/donate" button, since many apps on fdroid have a way to accept donations. You can scrape that from the standard metadata or the `index.xml`. I think it would be best to present it like https://elementary.io does: force the user to click "Buy" with a recommended value, but let them set any value including 0.
You can use this app to do it: https://f-droid.org/repository/browse/?fdid=de.j4velin.syste...
You can also disable the "install" script and have it install stuff without any prompts like the play store, just go into the settings, scroll to the bottom and enable "Expert" and "Install using system-permission".
Do you have/plan to build an installer app like the f-droid project? That app itself could use a modern design like you have done :-)