Focus on Business Apps
-
@scooke Just chill man. @eddowding is not alone in thinking that way.
Nobody has said that Cloudron would fail.
We are just business owners that have hope that Cloudron become bigger and add more business-oriented apps. We want their success and we are here to encourage them to pursuit that path.
If they speak out loud and simply tell us they are not interested to grow that much, fine. We'll respect their decisions.
But if they are, they have a big community behind them that is ready to continue submitting feedback and even agreed to pay an extra (if required) on their subscriptions to allow the Cloudron team to expand and add more business apps.
I have no idea why some people here try to defend Cloudron due to that post?! This is everything but negative.
I initially created this post to show my interest in having more business apps and many others agreed with me. This simply means that the interests are shared among multiple users and that my feedback should be considered seriously.
To those who are not interested, why are you losing your time posting here? I'm just not getting the point. Cloudron has many different audiences from businesses to personal. If you are not targeted by this, just move on.
-
imo the main issue in regards to new apps is the developer experience/on-boarding
Cloudron has had their own "standardized" method for packaging apps for many years now and have basically failed in facilitating a thriving marketplace for Open Source and new custom apps. Way before Supabase, Coolify and others "We"/Cloudron have had: Server/Runtime, Auth, DBs, Backups, DNS & so on readily available with absolutely no streamlined method to build an app using Add-On pieces and simply launch it.
This is the extent of it besides the docs
And don't get me started with the fact that we still don't have a single "no code/low code" APP/PAGE builder! No lie it makes me lose sleep at night...We can design a UI with Penpot, build databases with several options, build workflows with n8n, do AI stuff with Typebot/Open WebUI but I only have fucking wordpress to build a page!? (excuse my french & no offense WP but really!? it's basically 2025!!)
Ive been using Wappler over the past 4 years or so and even went as far as contracted a forum member Mitchelle to package the NodeJS app that Wappler outputs into a Cloudron package after trying and struggling for far longer than I'd like to admit. https://git.cloudron.io/murgero/wappler_app
Currently and Finally I have a decent tech stack that isn't missing a core/fundamental feature where I can have the assurance I need to build for and support clients but it could certainly be way more integrated than it is.Note: Wappler is not open source but does use an open stack and you own the output source*
There are plenty, maybe almost too many App/Page builders now to not have a single one on Cloudron: Budibase, NocoBase, Saltcron, Webstudio, Appsmith, and many more...shit, id even be mildly happy with GrapesJS at this point.Having met Girish in person and knowing the sheer engineering brilliance between you two I think the best path forward is to cultivate the current community and to streamline the ability for US to make an impact. For example making community positions or methods for us to contribute to Documentation, Press Releases, Video Content, Live Streams, 3rd Party Liaison, and IMO most importantly Education/Resources on Packaging Apps!
I know, I know before anyone says it! Sounds like a "skill issue" this is all just linux, docker and open standards, languages, protocols so you should know all of this already, its ASSumed that you do...and this is where the problem/current attitude toward "new apps" derives and resides.
Besides the Dev Exp the only other major issue is the lack of Docker Compose & Swarm for Multi Node implementation. This is the only other thing that also keeps me up at night. All this debating for years: To Kubernetes? or not? What about this hack with Distributed Replicated Storage System? Or maybe we roll or own? but that will take too long so no let's not do any of it. Meanwhile buddy over at Coolify wrote that functionality by himself, Twice!! And in half the amount of time we were talking about why we aren't using the most obvious implementation: DOCKER SWARM!!!
Its like three main functions: swarm init - take the addy/key and connect the server, update/redeploy. It's not a SpaceX mars mission guys and gals but we are officially late as fuck to the party.
https://forum.cloudron.io/topic/2990/3-node-deployment-unified-dashboardDon't mistake my passion for hate or anger
TLDR:
- App/Page Builder
- Better Docs/Dev Onboarding
- Compose & Swarm Implementation/Unified Dashboard
I pray this does not continue to fall on deaf ears...Especially since I am willing and able to contribute...
Everyone is here for a mutually beneficial relationship, let's keep that the focus or part ways amicably.✌💙+1
-
Yeah, that would be a good thing! I built my first custom app inside Cloudron and this has been a real puzzle... My dev has over 20 years of experience and due to the lack of documentation and no clear explanations, we needed to contact the support a few times. The support has been really helpful... Glad they were there otherwise we would still be probably stuck today.
A good example is n8n... What they are doing on Youtube is just amazing! Have a look at their Youtube channel, they have plenty of videos where they educate users and this has a lot of success.
I think Cloudron should do something like this... and even create a "community apps marketplace" where people can share their custom apps.
That would be greatly welcomed.
-
@plusone-nick said in Focus on Business Apps:
imo the main issue in regards to new apps is the developer experience/on-boarding
Cloudron has had their own "standardized" method for packaging apps for many years now and have basically failed in facilitating a thriving marketplace for Open Source and new custom apps. Way before Supabase, Coolify and others "We"/Cloudron have had: Server/Runtime, Auth, DBs, Backups, DNS & so on readily available with absolutely no streamlined method to build an app using Add-On pieces and simply launch it.
...And don't get me started with the fact that we still don't have a single "no code/low code" APP/PAGE builder! No lie it makes me lose sleep at night...We can design a UI with Penpot, build databases with several options, build workflows with n8n, do AI stuff with Typebot/Open WebUI but I only have fucking wordpress to build a page!? (excuse my french & no offense WP but really!? it's basically 2025!!)
...
I know, I know before anyone says it! Sounds like a "skill issue" this is all just linux, docker and open standards, languages, protocols so you should know all of this already, its ASSumed that you do...and this is where the problem/current attitude toward "new apps" derives and resides.
...
Don't mistake my passion for hate or angerTLDR:
- App/Page Builder
- Better Docs/Dev Onboarding
- Compose & Swarm Implementation/Unified Dashboard
I pray this does not continue to fall on deaf ears...Especially since I am willing and able to contribute...
Everyone is here for a mutually beneficial relationship, let's keep that the focus or part ways amicably.For points 1 and 2, I totally feel you. Anyway there is some hope and I want to share my experience below.
as an experienced developer and devops person, I have to admit it requires some skillset/experience and patience in order to be able to make the effort to package an app. As a busy parent, I had cooked a simple python app, very minimal, and packaged it for alpine, to boost my productivity. Then I wanted to host it on my Cloudron and it was not so easy to make it work, due to permissions issues, various errors (exec format etc), failed healthcheck... Despite having a working Dockerfile and docker-compose.yml in local, it took me 2 weeks of spare time to adapt my app for Cloudron. Mostly because my spare time is very limited and my knowledge of Cloudron packaging is limited too, and also I had a goal to only use ChatGPT.
The base Cloudron image is mostly ubuntu based while I wanted to keep alpine as a base image, and despite this difference I could manage to yet install my app on my Cloudron instance while still keeping it minimal
It took me a few ChatGPT prompts + a few evenings to go over all the issues I faced, yet I succeed! I believe the DevEx is quite horrible at this time, but with patience and help from ChatGPT
, and good documentation I'm convinced I could tackle any such challenge as long as the effort is realistic and the app is worth it. The more we cook app, the easier it becomes, and the funnier it is.I will try to share a bit about this effort soon
, so at least people can benefit from the learning. What is interesting, when building your own app, with a minimal setup, is that it is possible to avoid to rely too much on Cloudron "magic" as I don't care at all about the base image, etc.I had to admit that the information I needed to aggregate for the initial prompt to ChatGPT was huge and sourced from many different parts (docs, git...) and is not an effort everyone is willing to take, so it is worth cooking some ChatGPT app for this I believe.
Anyway the good news is that it is possible to have sort of a recipe to repeat this success, i.e by feeding ChatGPT with examples of Cloudron/Dockerfile/start.sh of similar apps; i.e apps made with similar tech stack; + relevant excerpts from the documentation about packaging apps for Cloudron and Cloudron requirements. And in total it took maybe a few hours of prompting to ChatGPT and git/docker push/cloudron install/update to make the whole thing work.
I hope to build maybe one or two more apps for Cloudron, for my own needs, this time I'll focus on apps I didn't build myself like maybe soulseek and others. Maybe in the end I'll have a good tutorial to help people contribute to Cloudron with sensible apps.
-
@SansGuidon good stuff i agree with the approach and sentiment
have you tried bolt.new or any other code assistant that can better generate? I know chatgpt recently added that canvas...but being able to have the ai be aware of the whole project context is right around the corner making this 10x easier. bolt.new has their OS version bolt.diy and can interface with all the usual llms, check it out if you haven't
This whole situation has basically forced me to be a better admin and learn to be a dev so cant really complain, just advocating for more facilitation✌💙+1
-
@plusone-nick Thanks for the tip on https://bolt.diy
I just built a game I've been wanting to build for the last 15 years using AI dev tools in an afternoon.
-
@robi said in Focus on Business Apps:
@plusone-nick Thanks for the tip on https://bolt.diy
I just built a game I've been wanting to build for the last 15 years using AI dev tools in an afternoon.
What's the game?
-
@jdaviescoates thanks for asking.. it's called Petals Around the Rose.
It's a great puzzle game one can play with a large group of people with just a handful of dice.
The innovation for my idea is the input method, instead of a separate text box and a button with modals that just distract from the game. It flows nicely with my design.
Try not to cheat ;-]
lol
