New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Will AI allow me to finally the end my reliance on WHMCS?
LosPollosHermanos
Member
I don't have an issue with them other than the gradually increasing fees making me feel like a frog in a pot of water on a stove, and I don't see any end to that trend. I've tried to create my own custom solution in the past but concluded it was just too much work.
Now it's a brave new world with AI tools. In only 2 days I have managed to create a partially functional customer management system. Frontend, Customer Portal, and Admin Panel. Still lots more work to do, but it would have probably taken me several months to get this far in the past. Looks promising so far.
Comments
some screenshots would be nice
Building a billing system can be complex, but it all depends on what you need. Plugins, payments, and so on can take a while to get right. AI will certainly help, but could also cause issues if you want to expand upon things down the line.
The most important thing with using AI is security; make sure you get that right from the get-go. If not, you could have someone in your codebase dumbing your db.
I would be thrilled to see a competitor to WHMCS.
I'll wait to hear about the data breach.
I'm trying
Off the shelf billing systems are more complicated because they need to satisfy lots of different requirements for lots of different people. A custom solution that only has the features I need is not necessarily that complicated.
Agree with you, that's why I said they can be complex, but it all depends on the business.
AI is way better at figuring that stuff out and finding the holes than most humans. I can also have multiple AI agents look at it, so I am not relying on just one training model.
There are already competitors and with AI they are all going to be getting much better and more competitive much quicker.
Just added a bunch of new features today, including multi-language support and a ticketing system. It's hard to believe that I can do stuff like this in hours now that would have easily taken me weeks or even months to do manually. It appears to be much more sophisticated and better quality than what I could have done myself as well.
I'm not an expert coder by any means but I am not bad. AI agents are much better and faster and constantly improving.
The problem is not building, but security. As soon as you launch it online and it attracts some traffic - automated scans will engage in hunt of vulns. And since vibe coding left you clueless about how application truly works and what is lacking. The edge cases.
You must write the concept your-self. No prompt will solve such complex problem as billing panel. Iterations upon iterations, bit by bit. No fancy tech stacks, just simplest well known tools. And only on existing codebase you can grow some llm augmentation.
Make it multiple nested layers. First, one instance of all top 20-30 models (check openrouter, it's cheap) review the code. Then combine this review into one file and pass it to the flagship models (latest version of Claude, GPT, and Gemini) to review. Then pass this review to Claude Opus 4.6 to summarize everything and fix. Now run this in a loop until all issues are fixed.
You don't know what issues actually exist unless you understand what you are doing. AI cannot be trusted to know what is and isn't an issue, because it will always have confidence in what it has generated itself. Trying to introduce more models just risks a feedback loop, which one do you trust to have the final say if you don't know yourself what is right and wrong.
Not a single word about trying to understand the way the pile of shit works, it's all in on glorified black box. Wow !!
I suppose some people don't need to understand nothing anymore, I don't know.
You can create anything using ai easily but should have some atleast junior level programming language knowledge
can you please let me know the name of your host so i can avoid it
AI definitely lowers the barrier for building something custom, especially for getting an MVP together quickly. That said, replacing WHMCS entirely is still a big commitment once billing, taxes, domain integrations, support workflows, security, and edge cases come into play.
Looks promising though. It will be interesting to see how far you can take it and whether it becomes easier to maintain than dealing with WHMCS over time.
We're slowly getting there..
Though it's not vibe-coded, I only tested the frontend-design skill - because I hate frontend dev. Not good, not bad... but until I'll have enough time to tackle that part myself, I'll have to live with it looking like 90% of AI sites.
Hey m8 I too hate frontend dev and anything remotely close to it ^_^ 👋
The backend part of VPS Price Tracker is so gay and colorful that internally the codebase refers to it as "frontend" too. The rule of thumb adopted by me is: anything not written in C++ is frontend, no matter if it runs on a basement workstation. Works for me :)
Why do you think hackers are using AI tools now? If you are not also doing that then you are the one who should be worried. At a bare minimum, you should be checking everything with AI tools looking for holes.
A custom solution does not need to replace all the features. Most people probably use less than 10% of the features in WHMCS.
Mythos is not yet released for public. Regular llm goes crazzy and want to refactor everything. No thanks. My twisted mind and creativity is what holds that damn app glued.
lol.
I usually refer to backend as any underlying layer that's not frontend...
For me this, for example, is backend:
https://i.ibb.co/SwY8y0YS/image.png
its the age of the vibe coded cpanels.
Forgot to mention... and got some DMs about it...
That's not a WHMCS replacement in a commercial way. Aka we won't sell it nor release it publicly. It's a project designed to drop our dependency on hostbill (as that's what we use now), because as awesome a panel could be, if that's ioncube encoded, it hardly limits what we can do, regardless if it's named WHMCS or Hostbill or any other name.
So, "SpanBill" is going to become our main billing panel in the future. Internal use only.
However, we do sell discounted BillingServ for our Webhosting Reseller plans customers! for those who are looking for a WHMCS alternative
I > @davide said:
I hate creating unit tests. AI agents are awesome for that. I just say "create units tests for the project" and presto. Then I tell it to add a generic agent directive to always add and run unit tests every time there are any changes or features added. The agent will then automatically run the tests and fix any errors it finds. That is one way to make AI agent coding more bulletproof and consistent.
Tests don't follow up changes in the flow - being maintained correctly and some times are even implemented wrong to begin with from AI . AI Definitely speeds up the process but tests must be reviewed from someone who understands
As opposed to lazy humans who often don't create any tests at all, often don't maintain them correctly, and often implement them wrong? Especially more experienced people who find that kind of coding really boring, and when they are under a deadline that will probably be one of the things they will ignore doing.
I can't imagine handling any big system without tests in place
Well... one of those systems was mentioned (imho) in the very first post in this thread unfortunately.