How I Built a Scaffolders Registry System Using ChatGPT (with zero coding skills, almost) ????

Ever wondered just how much you can achieve with AI tools for things you didn’t even know how to do? Like, really stretch the limits? Well, here’s my latest experiment: creating a fully functional online database system for a scaffolders registry – and I don’t have a single line of coding knowledge. ?? Buckle up for the full story of my ups and downs with ChatGPT, where typing prompts was my only “coding” skill. Spoiler alert: it was a wild ride! ??

The Challenge ??

Here’s the situation. I needed a system where potential employers or contractors could check scaffolders’ experience, certifications, and overall skills. Seems straightforward, right? Wrong!

So, what were my options? I could use a CMS that cost around €450 per month for maintenance, which might be “small change” for some, but that number adds up fast. Plus, it wasn’t even fully adapted for what I needed. Custom modifications? Extra cash ??. Consulting fees? Even more cash ????. And, of course, I’d still need to add specific features and pay for every tiny update.

So I thought, “What if I find a company to make a custom system for my needs?” In theory, it could be cheaper than a year of that expensive CMS (yes, that’s €5,400 annually ??). But, of course, updates, modifications, and maintenance would still cost extra. Nothing was quite the “perfect” solution.

And here’s the kicker: I almost don't know coding (almost) – fever PHP, CSS, or Java knowledge. To build my own system, I’d have to learn it all, and let’s be real, that could take years. So, what did I decide? Yep, to throw ChatGPT into the mix. I was determined to see if AI could help me build this thing from scratch.

Step 1: Choosing the CMS (aka, It’s WordPress Time) ??

First things first, I asked ChatGPT, “What CMS should I use?” After a bit of back and forth, it recommended WordPress. Simple, flexible, and (allegedly) friendly for beginners. So, with WordPress as the backbone, I was ready to dive in. Here we go!

Step 2: Hunting for Plugins ??

I figured, “There’s gotta be a plugin for scaffolders’ registries, right?” So I started hunting. ChatGPT helped me search, but every plugin seemed to lack one crucial feature or another. None of them really fit my needs. ??

Then, I had a “brilliant” thought: could ChatGPT create a plugin for me? It didn’t say “no,” so I gave it a shot. ChatGPT started spitting out code for functions like “add scaffolder,” “view certifications,” and “edit profile.” I tested it out with fingers crossed… but, yeah, the “edit” button didn’t work. Actually, none of it worked. ??

At this point, I was sending ChatGPT every bug I encountered. It kept giving me fixes – none of which fixed anything. Turns out, that creating plugins with AI isn’t as easy as I thought!

Step 3: Asking the AI for Real Advice ??

After giving up on the custom plugin idea, I went back to ChatGPT and explained everything I needed in painful detail. I gave it the full rundown on how the system should work, from ID searches to certification requirements. This time, ChatGPT suggested sticking with WordPress, adding a few plugins, and using custom PHP scripts for the tricky parts. Okay, not magic, but it was progress.

Step 4: The First Test ??

With ChatGPT’s instructions, I set everything up – scaffolders’ IDs, certifications, profiles, all of it. Ready for the first test! I typed in the ID, hit enter, and… Error. ?? Of course.

So, I tried sending screenshots of the error message to ChatGPT, and it spotted a missing line of code. It gave me a quick fix. I plugged it in, ran the test again, and… another crash. This thing was becoming my Everest.

Step 5: AI vs. the Wall ??

Here’s something to know about AI: if there’s a wall in front of it, it’ll walk into that wall every single time. It doesn’t “go around” or “find a way.” I explained each issue in detail to ChatGPT, and it would write code to “fix” the problem, even if that meant breaking everything else. ??

After about 10 rounds of back-and-forth fixes, ChatGPT finally produced a debug-friendly code snippet. At last, the system actually ran! It was messy and buggy, but I was getting somewhere. Small win, but hey – I’ll take it! ??

Step 6: Making It Look Good ??

So now, the system technically “worked,” but it looked terrible – like a 1999 website, and not in a retro-cool way.

So this is the first look:

The search results page was basic – too basic. I found some design samples and shared screenshots with ChatGPT, asking for a more polished look. It analyzed everything and tweaked the code, giving me a much sleeker, user-friendly layout. It was like magic! ?

I was amazed. After a few more tweaks, we had a decent design! ??

From here on, it was step-by-step improvements. I’d ask ChatGPT to tweak different visual elements, and each time, it delivered clear instructions and fresh code. Slowly but surely, the system was taking shape.

Step 7: Adding Fancy Features ?

Now, things were finally rolling, and I thought, “Why not add some extra features?”

• Expiry Notifications: I wanted the system to clearly show when certifications were expired. Boom – done. ?
• Automatic Email Reminders: Next, I thought it’d be cool if scaffolders got an email 30 days before their certifications expired. ChatGPT had me install SMTP and Cron plugins, wrote the script, and – success! ?? Well, sort of… it was so “excited” that it started spamming reminders every few minutes.

Lesson learned: be specific with your prompts. I told ChatGPT to only send reminders once. A few tests later, and finally, it worked perfectly. ??

Each little tweak added a new layer of detail to the email notifications. It started as a simple reminder, but then I kept asking ChatGPT to make it more personalized. First, I wanted it to include the scaffolders' names. Then I thought, “Why not add the card number?” That worked well, so next, I asked for the expiration date to be included, too. Each addition made the notification more informative, and with each prompt, the AI adjusted the code seamlessly.

This can be customized in any way...

Step 8: Adding Extra Functions ??

With the basics in place, I figured, “Why not go all out?” Here’s what I added:

1. Expiry Date Notifications: For expired certifications, instead of just showing the date, it now displays a bright “EXPIRED” label. ChatGPT handled it like a pro.
2. Automated Emails with Limits: I set up automated email notifications for scaffolders, reminding them of upcoming expiration dates 30 days out. Of course, ChatGPT initially made it send emails every time someone logged in, but after some debugging and extra prompts, it worked smoothly.

Alright, but how to keep those unique ID numbers safe so people can’t just guess others’ card numbers by entering random sequences? We could use a randomizer, Excel, or... ask ChatGPT again! So I asked, if we have a field named "card_id_number," can we auto-generate something like the first two letters followed by seven random digits each time an admin enters data? And voilà, ChatGPT came through with the solution!

The Result ??

But here we are—a fully functional scaffolder registry system built from scratch, without hiring a developer or learning to code. This project, which might’ve cost thousands in development fees and consulting, was built with nothing more than ChatGPT and a bit of determination.

In total, the setup took about 8-10 hours. Imagine what you could create with days or weeks of learning prompts! So, if you’re wondering if it’s possible to build a working system without knowing a single line of code, the answer is yes! Just get ready for a few crashes, lots of re-prompting, and an adventure with your AI buddy that’s sure to keep you on your toes.