Show HN: A searchable list of self-hosted software with screenshots
selfhostedsource.tech
selfhostedsource.tech
I fantasize about running a little consultancy that would set up and maintain a tailored package of self-hosted OSS software for such small businesses. But I haven't actually studied whether there's a workable business model to be had, or whether there's enough quality self-hosted software out there to adequately cover the needs of most small businesses. I'm curious if HN thinks this could be a viable business...
They raised $250m a couple years ago, and have been acquiring smaller companies like theme shops, local site management tool based on Docker and Electron, etc. They've got a mature platform suitable for businesses, and it looks like they have a very healthy growth trajectory.
On the other hand, for me they're becoming the "big company" in your description, and I'd prefer a self-hosted or at least open-source solution.
Thanks for the effort you put into simplifying the use of all those applications!
A disclosure reveals some previously unknown information, while a disclaimer renounces your responsibility for something.
So your example is correct, although "I am not a lawyer" could also be considered a disclosure.
Disclaimer means refusing to accept responsibility for someone else‘a decision to rely on you. When people are afraid of getting sued, they disclaim liability.
Shift/inventory management would need to be added via third-party software (or implemented in spreadsheets).
I've heard of companies setting these up for small businesses. [5]
[0] https://www.synology.com/en-global/dsm/feature/mailplus
[1] https://www.synology.com/en-global/dsm/feature/calendar
[2] https://www.synology.com/en-global/dsm/feature/drive
[3] https://www.synology.com/en-global/dsm/feature/chat
[4] https://www.synology.com/en-global/dsm/feature/office
[5] https://www.synology.com/en-global/wheretobuy/United%20Kingd...
So there may be a business opportunity, but you need enough scale that each single customer is happy with the price.
I believe there is a niche to be served here, just not sure exactly how yet.
Also, I once saw a "Linux-friendly" PC shop, literally sign-posted "LINUX SHOP", out in farm-boy country, in the lower Austria sprawl, so.. pretty sure there is a mom 'n pop/cousin shop mentality for a lot of IT needs...
The OSS is out there, but you've gotta wrap it in enough design work to make the experience comparable to closed source offerings: Most businesses aren't going to want to pay for worse software.
Disclaimer: I am the co-founder
For instance, I run my own Nextcloud server for 10$/month, the amount of maintenance per month is 0 hours. Nextcloud has already provided a very simple installation setup using snap. It would be cheaper, in the long run, for someone to pay a contractor 300$ to setup a 10$/month droplet on Digital Ocean.
What I'm getting at is you may want to consider per app pricing as a customer is probably interested in a very specific app and thus will price check only for that app.
We have so far not wanted to get into per-app pricing, because we see ourselves as app packagers and not providing support for the app itself. For example, we don't actually provide any support if Rocket.Chat's mobile app has some problem. Making it per-app pricing can mislead people into thinking we support the app. There is also the issue that we might be seen as competing with the app authors (if they have a SaaS model).
The 30 USD is essentially the price for automatic updates, backups, dns/tls management, deployment with security best practices etc. It is also a 'service' where you can contact us if something goes wrong. As you say, if you want to DIY/time is cheap, cloudron is expensive. That said, if you use the DO marketplace image, you get a 50% discount (you will see the discount coupon when setting up a subscription).
Anything involving the command line is considered advanced knowledge for most people, so targeting your product to the novice is a miss.
the command line. a beginner is not going to open the command line and start typing in commands.
"If that's the case, we do have Cloudron listed in the DO Marketplace"
I would suggest break this page (https://cloudron.io/get.html) into a beginner and advanced section and give a detailed instructional walk through (with pictures) for beginners.
...and some animation of Mickey Mouse & cie congratulating the user with some school notation at every step of the tutorial. Like this :
"Great, you didn't forget to prefix your command with 'sudo' while avoiding the evil 'root' user. You deserve a A+. Now take a candy bar in the bag provided by Cloudron just for you."
Most people could probably understand the technical stuff if they had enough spare time/interest, but if you don't have the technical literacy for it, "How to install on Ubuntu Bionic 18.04 x64" is confusing.
Whereas, "I followed these steps on this page (and the screenshots are how it looked like)" is easier.
Let me know if you’re interested in onboarding it into Cloudron :)
The thing I haven't figured out is what happens to it if/when Cloudron goes away. Do I have/get the passwords somehow so I can hire someone else to take it over (or, heaven forbid, dive in there myself)?
We won't be factoring privacy, working with a big company, or other concerns in this thought. We will only focus on the numbers.
Lets look at email/productivity market real quick.
Google's base package is 6 dollars a month per user with all the backups, support, and infrastructure in place (realistically, you'll want the 12 dollars a month package).
Since email is critical infrastructure (I believe it's one of the most critical elements of a business), lets say two DigitalOcean VPSes at 6 dollars a month each (one primary, one failover) with a license of HostinaBox/Vesta/Whatever Open Source Solution you use with DigitalOcean's backups enabled. Not the "best" design but something I'd say is OK and within scope for a small business. That's 12 dollars a month base costs + your consulting time for a sizeable capacity.
For a small business of 1 person, they get a better deal by just going with Google. Google has their applications and softwares easily integrated with their other services, comes with their productivity suite (GSuite == Google Docs, Drive, etc.), and as a small business who is probably risk adverse in their decisions when it comes to these things, feel more comfortable working with Google.
For a small business of 5 people, I'd say it's still more worth it for them to use Google as that's 30 dollars a month (most consultants charge more than that an hour). If they hire a consultant and if the poo hits the fan, then they'll be paying a consultant money to execute the disaster recovery plan. Even if you did take them on as a client, that's a maximum of 18 dollars a month you get to keep (assuming no issues/errors happen).
For a small business of 50 people, then now it gets to an interesting territory. However, for 50 people I'd change up the base server/system configuration to have higher capacity, more fault tolerant, and resilient under disaster scenarios (which would increase base operating costs).
I'd say this really depends on marginal benefits and based really on relationships established with your clients. In the end, you can probably make some $$$ but your time and effort might be spent on more productive and lucritive tasks. This is also assuming that the self-hosted OSS software is of quality that the clients will be happy with. I'd argue Google's mail offering may have annoying/restrictive spam policies and be frustrating at times, but they have a high quality product made at an affordable price point. The variation in quality of OSS products concerns me as well as developers who are probably overworked and underpaid for their contributions asked to make changes to support clients they're not directly paid by.
As a risk averse business, I'd rather rest my eggs in the Google/Microsoft/whoever basket and directly work with the "entity" that maintains the codebase (or has the talent/expertise on-hand to make adjustments) rather than a middle intermediary of equal level but, in the end, is subject to the decisions and leadership of the OSS product.
Now take this to another step and say you build a consulting company that handles all of these as a one-stop-shop? Well... Then I don't see anything new service/model here than what a local IT consultant company can offer.
So to really make this work I'd say a shop that automates these deployments on-demand and offers a large selection of applications to use is probably the best step forward. Even then, I don't really see the viability of this on a funding perspective except for scaling. In the end in my perspective, the opportunity is there but it'd be hard to do it right. Also funding the developers of the software you're making money off of would be great, but that's a whole nother thing (and I can squabble about that for hours).
Quick plug: two very good friends of mine are in the process of tackling a similar issue via their venture[0]. I am a customer of theirs but have been friends with them even before this venture. Really recommend their product as an affordable and reliable product that "just works".
[0]: https://mxroute.com/
Heck, I personally "mantain" (not that I do much), all my company employees email address, different domains, a couple of simple webpages, calendars, issue tracker...
All without doing much apart from setting the anti-spam, and installing the Installatron applications I need most.
I wanna write a book titled: "how to run your small/medium-sized office with just a linux box on the wall and your staffs cell phones..."
Pretty sure I read a FAQ for how to do this, 20 years ago, though...
> Lucid index sources it's data from curated lists of software compiled by volunteers. The specific lists used are:
> - Awesome Self Hosted[1]
> - Awesome SysAdmin[2]
I think it would be a huge improvement to give me a sense of whether the product is well supported.
There are a few reasons for this...
1. I want to choose an OSS that will have security vulnerabilities resolved in a timely fashion.
2. I want the product to keep up with the times and changing landscape.
3. I want to be able to converse usefully in a forum to resolve issues.
4. I don't want to search for a new replacement product and undergo a migration because all the devs have abandoned the project in x years.
That said, the layout of your site is very nice! How does one get new projects added to the list? (For example, we would like to have Cloudron itself listed there).
In principle, I'm a big advocate of self-hosting (one of my services is even on this list). In practice, it just doesn't work for me. Once I get beyond 2-3 services it's just too much hassle to keep track of everything.
The key realization for me is that I don't actually care too much where the software is running, or who is running it for me. What I do care about is a avoiding vendor lock-in. As long as I have a reasonable escape hatch if my service company starts doing things I don't like, that's good enough. This keeps them honest. My issue with the current crop of monoliths like Google services is that there's no obvious migration path if you get fed up with them, so you're pretty much stuck with them no matter how crappy their software or customer service is/gets.
That's why I think something like sandstorm.io or cloudron is the future of self-hosting, at least in the near future. Maybe eventually we'll have a substrate of simple protocols and practices that will make it reasonable to manage everything yourself, but we're not there yet.
I used to install and upgrade everything myself until I discovered docker and docker hub.
Together with watchtower I have zero maintenance on the versions and everything updates automatically.
Docker Swarm shines in this regard because it's easy to setup, and its simplicity allows you to just forget about it. I run 30+ services behind a reverse proxy for personal use and don't need to keep track of anything.
Shameless plug: Would you be open to listing our project Rudder, an open-source Segment alternative. https://github.com/rudderlabs/rudder-server/. We were on HN recently
At the moment people, myself included, use "self-host" for both.
Whenever you expect to get a traffic spike, you can always turn proxying on for a day or two and turn it off afterwards.
I'll have to agree, there are few but very vocal.
https://selfhostedsource.tech/self-hosted/?page=2&search=web
(Not sure if a bug or due to load.)
Am I blind or there is no 'Submit' section on the website?
https://selfhostedsource.tech/self-hosted?search=heimdall&li...
Then I noticed under filter was categories.
So, actually, its perfect.
Am I meant to put this software on a computer and run it 24/7? Why?
It's not 'technically' part of selfhostedsource, but there's one database shared between selfhostedsource and lucidindex so it's pretty easy to get results from other lists
Lucid Index will eventually index all the 'awesome' lists, but there are a few others already functioning here: https://lucidindex.com
Beyond that, I don't really have any specific plans to monetize.
Search could use some work:
- Search for Bot, get a bunch of unrelated results.
- Search for Email, get accounting/translation/etc tools.
- Search for Accounting, seems to work correctly.
Otherwise my stack is completely self made and can run on a Raspberry: http://github.com/tinspin
Could you elaborate what that is supposed to mean?
Would you be willing to add my project (10K+ GitHub stars)?
It is a self-hosted Open Source alternative to Firebase, with graph features, and fully decentralized. It is used in production by HackerNoon and other sites with millions of users.