Self hosted FLOSS fitness/workout tracker | Hacker News Reader

Self hosted FLOSS fitness/workout tracker

github.com

221 pointsfreemh112 comments

This app seems to require a server. Invoke the app, and it presents a login screen, with links to "register" and to "use custom server".

I suppose an alternative design would be to be on-device storage first, and then have an optional sync to a server (or laptop/desktop).

Normally when talking about fitness tracking we are talking about embedded systems with very limited capacity both in terms of compute power and storage. You can use on-device storage to buffer tracking data, but any relevant and long term assessment and storage of the data has to be done off-site.

This app is not about biometric fitness tracking as in smart watches, it is about manually tracking workout information. In that case, there is no need for a server to start with.

Huh, I totally don't get this conjecture... What do you consider fitness tracking? In my mind a basic fitness app is essentially a replacement for a journal.

It's an Android/iOS app. On-device storage is measured in gigabytes, on-device CPU is very capable.

They are commenting about smaller devices such as a smart ring or a smart watch. Running a web server on such devices is the best way to deplete their battery.

Seems relevant to the blog post I just read yesterday!

https://raphael.lullis.net/thinking-heads-are-not-in-the-clo...

And what was posted yesterday on HN as well.

https://news.ycombinator.com/item?id=43021677

Indeed, this should be local first software [0], not self-hosted but server required.

[0] localfirstweb.dev

Generally firebase / supabase do a very good job of providing client sdks to access database directly without the need of a server.

That way your data is still in cloud and you don't actually need a server.

Really? They do not run on servers? Where us the data stored then?

This is unrelated to local first, just because you don't need to set up your own backend (not a server, because they are all servers) does not mean that the app works purely on the client.

Supabase and firebase are servers.

In fact if something is "in the cloud", it's by definition being made available by a remote machine, AKA a server serving the data.

The alternative is not to have a server and instead have clients share data if data needs to be shared.

True, one of the pointers in the comments was that people don't want to host their own servers so was suggesting an alternative to that.

That's even worse. The whole point of things like this are you do all your computing with your own computer(s). People in this thread are just saying that it would be even better if you could only use your phone/laptop etc and didn't have to also set up your own server.

That's what "self hosted" usually means.

To be fair, the app being posted is the server application which you can run locally. You do not need the phone application to use the web application.

The mobile app is also open source https://github.com/wger-project/flutter

I don't know how but I imagine you could build it and side load your own build to your phone?

Yes, that's what "Self hosted" means, you install it on your own server. Your choice if it's something "cloud" based and remote from you, or resides in your home.

https://wger.readthedocs.io/en/latest/installation.html

Is there a missing "product" for [easy] "self-hosting" in the cloud? (Or does it exist? Or will it only apply to a narrow kind of user?)

Like I think there should be some way to "one-button-click" install "self-hosted" apps in the cloud, tied to my personal account (and maybe with auth tied to that account). And I pay the usage fees for the cloud (hopefully on a per-request kind of basis, not an always-on server instance).

Is this a thing that I don't know about? Or is the market too narrow to be useful? (Otherwise, why doesn't it already exist?)

I think the most popular service like this is PikaPods.

I hadn't heard of it before. Thanks!

The problem is that self hosted apps are rarely designed to be run serverless (why would they be?) and giving each app it’a own VPS or hosted container is going to price out the self-hosted crowd, to the point where you might as well be paying for some cloud software.

In particular, self hosted apps usually are using relational databases or SQLite which need persistent disk so can’t run serverless. They also sometimes require writing to physical disk instead of object storage like S3. Writing or rewriting apps to support serverless when they have no technical need to when self hosting would make things more complicated. Most CRUD frameworks used to write self-hosted apps do not work with NoSQL out of the box.

Thing is, almost every self hosted app supports docker now and so if you like, install portainer on a VPS or NUC or raspberry pi and you’ll be able to set up most self hosted apps easily without touching the command line.

> and giving each app it’a own VPS or hosted container is going to price out the self-hosted crowd,

As far as I know, myself and other self-hosters run these sort of applications/services on home infrastructure or VPSes/dedicated/bare-metal where multiple applications usually share one instance. This could be done with docker, or cgroups, or countless other ways. I'm not sure if that's what you mean with a "hosted container" though, don't think I've heard about that before.

Yes. But that is not what OP comment is asking for. They want one-click. And request based pricing. I was explaining why request based pricing is infeasible and one-click install would price people out (because it would imply a VPS per service).

And I said the same thing at the end of my comment about the way people would host things using docker on a VPS or home server.

> Thing is, almost every self hosted app supports docker now and so if you like, install portainer on a VPS or NUC or raspberry pi and you’ll be able to set up most self hosted apps easily without touching the command line.

Why would you use a new VPS for each application? Just dockerize and put all your apps on one VPS, which is what most self-hosters do.

Because that’s not one-click setup or priced per request, which was the comment I was responding to was seeking.

And I did say at the end of my comment:

> Is this a thing that I don't know about? Or is the market too narrow to be useful? (Otherwise, why doesn't it already exist?)

I'm not super familiar with it, but I think what you describe was/is the goal of Sandstorm (https://sandstorm.org/).

Then there are also efforts like YunoHost (https://yunohost.org/) which are kind of like that, gives SSO auth and everything out of the box for all the apps it supports.

Even easier to use and less involved would be maybe what TrueNAS Core has in terms of apps support, which is essentially also "one-click install" of self-hosted applications, backed by local Kubernetes installation if I remember correctly.

A more involved option for people who want to manage more themselves (both infrastructure and configuration-wise) is using NixOS, which is the approach I chose for my own local infrastructure at home. For the packages supported by NixPkgs, many applications are like ~4 lines of configuration to setup and get integrated with the rest of the apps you run.

All of these options you can run in the cloud, bare-metal or at home servers, afaik.

They are now backed by docker after the Electric Eel update.

Had to look it up, you're talking about TrueNAS moving from Kubernetes to Docker, just for clarification, https://forums.truenas.com/t/the-future-of-electric-eel-and-...

Huh, interesting as I literally moved from TrueNAS to NixOS because of two reasons 1) not liking something as complicated as Kubernetes for something as simple as home infrastructure and 2) to have a more reproducible setup.

Happy to hear 1 won't be an issue for others in the future :)

I feel that, I went back to truenas for the simplicity.

This 5 disk pool started from the freebsd truenas, to Debian, to nixos, back to linux truenas.

And the disks where bumped from 2TB to 12Tb

Not sure if there is something like that tracker but I think you would like something like Linode marketplace:

https://www.linode.com/marketplace/apps/

I used to hear that Yunohost was a good option, but I am not certain if this is still used by selfhosters today.

I was thinking of building a self-hosted app framework that manages the application lifecycle like how smartphones do.

The only problem is that the existing apps wouldn't work with it.

Cloudron solves this going from the other side: they package apps themselves. To support transparent migrations and backups.

Ryot is an 'everything' tracker and has a good section for workouts. Very easy to deploy.

Look for self-hosted PaaS, platform as a service. Coolify and Dokploy are good options, on Hetzner.

Yes! There are lots!

Even decades ago, you could buy a web hosting account and simply click an icon to install Wordpress, CRMs, webmail clients, etc in your account and get started with minimal hassle. There are very likely many of these still around. Of course, if you are not a technical user, you are limited to what they provide.

In the realm of containers, there are also many many choices for this. Most are open source, some are commercial. The problem with all that I know of, is that when you want to use an app that isn't in their "store", or when you want to use it differently than how they have packaged it, either you can't do it because it's not supported, or you essentially have to learn docker from scratch anyway.

Shared hosting providers have had such installers, typically as part of cPanel, for decades. An example would be Fantastico https://www.wikipedia.org/wiki/Fantastico_(web_hosting)

You’re not going to get per-request billing, but shared hosting is typically as cheap as you’re going to get for “always-on” web hosting.

Aren't most people interested in self hosting specifically to get away from the cloud?

I don't know if you can reasonably generalize self hosting.

To a lot of people, self hosting is about getting complete (or effectively complete) control of your data and privacy. That can be compatible with the cloud.

Yes - and/or data sovereignty.

Several apps that try to solve this. Good one, but costly, is cloudron. It allows you to one click install a wide range of self-hosted apps. Cloudron itself is subscription software that can be installed on own server or many cloud options.

There are many more, though. Things like proxmox with app stores, which are easily installed in most clouds.

Also CasaOs and others.

Another interesting self-hosted tracker is https://ryot.io/.

Demo: https://demo.ryot.io/_s/acl_vUMPnPirkHlT

Code: https://github.com/IgnisDa/ryot

(PS: they opted to go in a direction where it also includes some media-related features -- like tracking movies or books you've seen/read -- but this can be disabled).

Neat project, thanks! But, I gotta ask... what does the "L" in FLOSS stand for? :-) (I've only ever heard of FOSS.)

I believe it means libre or free as in free speech, not free as in free beer.

Libre.

See https://en.m.wikipedia.org/wiki/Alternative_terms_for_free_s...

Usually the L stands for Libre indicating full freedom of usage of the software and that is not just "free" as a price tag.

https://www.gnu.org/philosophy/floss-and-foss.en.html

Ok, this one’s baffling, especially learning at the end that FLOSS is bad because it’s meant to de-emphasize the “libre” part.

All OSS software is inherently without cost, that seems unquestioned here. So free only ever means one thing to non-laypeople, in this context. So isn’t FOSS already the neutral middle ground between OSS and FS??

Regardless, I’m struggling to conceive of how a piece of software could be OS but not F. I guess if it’s, like, surveillance software known to be used by governments…? Maybe OSS that is paradoxically restrictively licensed, threatening any forks or unauthorized compilations with legal action? That seems like a terribly naive proposition, but I’m sure it’s been floated by at least one MBA…

In other words: you can argue all day about the justifications for OS’ing your S being more related to removing cost barriers or to sharing control, but in the end, you clearly have to do both. Making “F[L]OSS” redundant at best, confusing at worst!

Surely I’m missing something, bc I know this has been litigated for many thousands of hours both pre- and post-Eternal September. But rn it just comes across as baseless pedantry

> All OSS software is inherently without cost, that seems unquestioned here.

I don't think so, you could charge money for FOSS (like charging for a built binary but having the source be FOSS) and it'll still be as much FOSS as any other FOSS out there. It isn't very common to do so, but there isn't any inherently wrong or incorrect with charging for FOSS.

How could you charge for a binary if people can just compile it on their own...? Honor system? I guess you could make it inconvenient to compile, but then is that really OSS?

> How could you charge for a binary if people can just compile it on their own...?

You don't make a binary publicly available, then you put the binary behind a paywall.

Some examples:

- Ardour - Lets you pull down the source and compile locally for free, or you pay to use their compiled binaries. Author/creator of Ardour hangs around on HN, maybe they could share their experience if they see this.

- Radium - Another DAW like Ardour, does it the same way.

- Fritzing - Designer for PCBs, same approach, pay for the binaries if you'd like, but free to compile from source if you can

I'm sure there are many more examples out there, but these are the ones I thought about when I wrote my previous comment.

It's interesting to reflect on what you're saying: you'd pay someone who forces you to, but not otherwise. So if someone built a house for you you'd only pay them if they threatened to come and burn it down, or they kept some way to remotely lock you out of it?

Hmmm. Well it’s software/IP, so as always I think we need to stay away from “would you download a car” talk.

With that in mind, the proposition is basically just the honor system. Which maybe works a little sometimes, among professionals? I paid for SublimeText to support them, for example. But WinRar is a very compelling counter example.

It feels like publishing a pdf of a book but with a big red “don’t click this until you’ve Venmo’d me $5!” above it. Regardless of what I individually would do, that’s just kinda… goofy?

Initially there was no Open Source or OSS. There was Free Software. This started a long time ago with things like GNU and later Linux, BSD etc. Richard Stallman codified what it meant for software to be Free. Specifically, it had nothing to do with cost. The "free" was free as in freedom, not free as in beer. Unfortunately English uses one word for both senses. Romance languages like Spanish still retain both: gratis means free as in beer and libre means free as in freedom.

Open Source was a later "rebranding" of free software by some people who fell out with Stallman and wanted to emphasise more the practical advantages over the ethical ones. Stallman wasn't happy because he felt (and still feels) the most important thing is that each person should be able to do computing freely.

Anyway, long story short, free software is nothing to do with cost. Horrible acronyms like "FLOSS" are to try to make everyone happy.

The "F" (free), "L" (libre), and "OS" (open source) all mean the same thing, the acronym is just meant to make extra-sure to include all the ways people refer to that kind of software (last "S).

They absolutely do NOT mean the same thing.

There is open source software that requires payment to obtain.

There is open source software that you are not free to do whatever you want with (non-libre)

There is software available at no cost which is neither open source or libre.

Just about every permutation of these 3 concepts has some actual example in the real world. Please do not promote the idea that they are in any way synonyms.

In this case they are basically synonyms, the Free in FLOSS doesn't mean free of charge and open source doesn't mean source available.

What is an example of open source software that is not libre?

Like the other commenter said, the F in F(L)OSS does mean free as in speech, not necessarily free as in beer.

People disagree about this. If F is for free as in speech, what is the L for?

I've been in the GNU/libre/open source world since 1986. My experience is that the L (for libre) was added specifically because the english "Free" is ambiguous and too often interpreted as without-cost/free-as-in-beer/gratis.

Libre more specifically means the GNU licenses, as GNU didn't like that licenses like MIT were being called FOSS, as they don't preserve user freedoms like (A|L)GPL do.

This is absolutely not the history I have lived through.

https://www.gnu.org/philosophy/floss-and-foss.en.html

Stallman wants people to say FLOSS to denote free software versus open source software, and my interpretation of this article is that he only deems the GNU licenses to be truly free (libre) software.

Stallman acknowledges the he did not invent the term, and describes how the "L" was added to connote freedom in the free-as-in-speech sense, rather than type of freedom the "F" was typically associated with.

FOSS was already a term to denote the other types of open source license like MIT, that is why FLOSS was specifically coined, to not only say that there is a difference in the word free, but also to differentiate GNU OSS license from the others.

> [..] only deems the GNU licenses to be truly free (libre) software.

Simply has never been the case.

> What is an example of open source software that is not libre?

One example I am familiar with is an (old) project called LinuxSampler. It was (probably incorrectly) licensed GPL+modification clause that prevented its use in any commercial hardware project.

There have been some other projects that are released under what might be termed "GPL-but-no-military-use".

Then they're not really open source then? Incorrectly licensing them doesn't actually mean they're open source.

Unless you insist that open source means "there can be absolutely no limits whatsoever on the use of the source code" (which is not an unreasonable position), then they still fit into many people's sense of open source - they can do anything they would want to do with it.

The people I know who have added clauses like this do not believe they are breaking the spirit of open source; I'm not here to insist that they are right or wrong.

I'm here to insist they are wrong, because people's notion of a term is not actually what that term might mean. Call me an etymological prescriptivist but OSS has a defined meaning that no amount of vibes-based hand wringing will change, it pertains to the OSI definition in my view.

And let's not forget, licenses are legal contracts, so it behooves people even more to get them correct lest they be sued.

The terms "OSS" and "open source" do not appear in the licenses associated with them.

Why would they be, literally? That doesn't mean anything about whether they are in the category of open source or not.

The “NASA Open Source Agreement 1.3” is an Open Source license that is not a Free Software license (hence not Libre).

There is free software (as of FSF) that require payment to obtain.

Are there any privacy-friendly wearables in the fitness tech scene? It seems like most wearable gadgets send data to remote servers, whenever possible.

If you just need heart rate, you can pull data off a plain sensor to your phone or to a computer, they report over BLE.

What sensor would you recommend? I have tried Apple Watch and i really only care about the heart rate. But it doesn't poll the heart rate constantly unless in exercise mode. I wish I could just constantly track the heart rate every second.

I use the basic garmin chest strap. It transmits continuously and supports things like HRV, which I suspect is a secondary high data rate mode.

Most Garmin watches _can_ work without connecting to your phone and the app. It just stores the files locally on the watch and you can just connect it to your PC and read the files.

Please let me know if you know of any options to streamline the process.

Bangle.js 2: https://banglejs.com/

Depends which sport - but golden cheetah lets you link a range of apps to your PC and store locally

I got to know more at your stand during last year's FOSDEM. Very cool project!

I'm using it (with the hosted service and app installed from f-droid) and we'll, the UI could be improved. Bigger issue is that there is no way to know in which series you are at the moment? Is the 3rd? The 4th? The 5th? Also the timer between series stops working if you put wger in the foreground (I disabled all the battery optimizations etc for it in Android 15 on a Pixel 6a). Also some features are just in the web UI. But overall, it's a nice system and I keep using it.

Have you considered sending a PR? :)

Yes I did, but I'm not very skilled with mobile development and I don't want to take time from the gym ;) Anyway I'm going to have a few spare days in the near future, I might try

Excellent, will tty it out and report

That TTY typo reveals you're a true nerd ;-)

I tried to use it but I found the UI/UX confusing, with many exercises lacking images too. I support the idea but the workflow was not for me.

I settled with Iron for iOS, which is also free software despite the platform it runs. Still looking for an equivalent on Android, the LiftLog app seems to have potential, but I think the UI is not that great.

I like Liftosaur, which appears to have an Android version on the Play store.

Tried a little bit, it seems nice, but it is a bity janky on iOS. Feels like I'm using a web app, and some features require subscription/account.

i use fitnotes on android, it's not foss but there is no tracking and absolutely no internet connection if you don't enable cloud backup.

the ui is old but effective.

I've been lifting for ten years thereabouts. Used JEFIT, Hevy, and Strong throughout that time.

I've given up on apps completely and log with a reMarkable.

All of these apps use their own data schemas, which makes normalizing the data for export into something like Google Sheets really hard. Consequently, you're locked into those apps and whatever design choices they choose to implement.

Moreover, I don't like using my phone when I'm at the gym, and, well, they're phone apps.

For strength training, I can recommend https://www.liftosaur.com/. UI is a bit clunky, but very powerful, it even has scripting language for describing workout plan.

UI is bit clunky. If we had great UI for this app, even if not open source, then lot of athletes and runners will use it.

Can I try with a hosted version first?

Stepping back a bit, we used to call these "programs". Why are they now "hosted"? I see this with recipe databases, fitness trackers, all sorts of things that have no reason to ever leave my laptop except when my backup job runs.

What's the advantage of having the server and data on some device other than the one I'm using to interact with it?

My impression is that “hosted” is common parlance for programs that present as web applications, rather than typical executables. A web application is naturally capable of “leaving the laptop” so to speak, which is often one of the benefits/intents of the program design.

Because hosting it on a server allows for automatic data entry via a mobile phone app?

No a lot of people would use fitness trackers if you manually had to enter all the information.

For fitness trackers, you're often going to get sensor files directly from a wearable device, but that device may have limited or even no interactive IO facilities, making it unsuitable for planning and analysis, which requires exporting to a more capable device. That device may very well just be your laptop, but you'd still want the ability to directly receive sensor data that your laptop can't collect because you can't hook your laptop up to a heartrate monitor and accelerometer and GPS while you're working out.

For a recipe database, if you're thinking of something like a food logger, having the food database shared by all users instead of local to each user is what allows the data to already be there the first time you ever log that same food. If it has a barcode or simple description and anyone else has ever input the nutrition data associated with it, it's already available to you without every user having to input that data individually.

Inputting data on your smartphone and it appearing on your laptop without any extra sync effort.

I like to think of it not as "computer program with extra steps" but rather "service I can access from anywhere/anytime using any device with added privacy and control".

DISCLAIMER: I'm a backend software engineer so I'm biased.

100% agree

All replies focus on "service" or cloud-in-general is needed to implement sync between devices.

Why we don't have "own" sync-my-files as-service that can be connected to many apps?

I would prefer this one (in all incarnations - desktop/mobile/watch) to be just apps, but connected to something like "folder in dropbox" or "folder in gdrive" ...

My cloud my choice. Their program my data (here or in my cloud).

I know that it doesn't work in todays economy when everyone expects free in exchange for some invonvenience like ads/data-scraping by vendor.

>apps, but connected to something like "folder in dropbox" or "folder in gdrive" ...

That's the opposite of gp's question of: >no reason to ever leave my laptop [...] What's the advantage [...] data on some device other than the one I'm using to interact with it?

Gp was wondering why the data can't remain local to that device [i.e. laptop]. Proposing "choose-your-own-cloud-service" like DropBox and Google Drive contradicts that.

> What's the advantage of having the server and data on some device other than the one I'm using to interact with it?

Mostly that many people have more than 1 computer these days: a small one that just about fits in the palm of their hand, and a larger one that stays in a room.

This means that most people are instantly operating within the realm of "distributed computing", and one of the fundamental approaches to that sort of thing is to allow one of the computers to both provide and accept data to/from the other.

Of course, if you have an application that will only ever be used on a single computing device, and whose data has no utility anywhere else, there's no need for the data to be able to move to/from a server. Such applications do exist, but so do many, many more where the ability to view and modify the data on different devices in different locations is desirable, even when there's just a single user involved.

Until and unless storing data locally is your USP, it can create issues as well for the developer

- They might have less control over the algorithms deployed on client side vs them being on the server. In case of android apps it is a hassle to get the users to upgrade the app. - In case there is a bug in the data I can't easily see the shape of the data and debug faster.

Probably in this case developer might be ambitious and would be thinking about adding extra features to their stack which would be potentially easy if algorithms and data is under their control.

To make it super easy to use the same "program" on your smartphone/tablet/laptop/desktop, wherever you are. And/or to simultaneously share access to it with friends/family. Yes, you can all just point a local app at a data share instead (the old way), but many apps aren't available across all these platforms to even make this possible across iOS, OSX, Linux, Android, and Windows, so this often doesn't work for a group of more than two people.

> many apps aren't available across all these platforms to even make this possible across iOS, OSX, Linux, Android, and Windows,

Thanks to all the spying/telemetry in most of those platforms once you use your privacy protecting app on them you lose the advantage of your data being self-hosted in the first place.

Can’t import from Apple Health. Looks promising but this is a major draw back.

build it? It's foss

https://github.com/wger-project/flutter/issues/618

Reply on news.ycombinator.com