Adventures With Vibe Coding: Building a SaaS Platform - Day 1

Vibe coding hype is real. I don't think there is a single moment when I look at any social media when I'm not being attacked by Vibe coding. It's a new block, it's a new gold rush - the hype that everyone can build an application in 30 minutes.

Even though I've been all the time skeptical to it and I even wrote about what it means for Khiliad here , I decided I will give it a go and see what happens. This is going to be a growing article, I will be adding my experience day by day, and we'll see how it goes.

Rules of The Game - I Know Nothing About Technology

In this experiment, I will pretend I don't know anything about software development, technology, building SaaS platforms or anything related to that.

I will play complete ignorance. I’m going to walk through building a real SaaS product using AI tools, acting as if I have no development experience at all.

I won’t read documentation. I won’t write or understand code. I’ll follow AI prompts blindly. I’ll ask stupid questions. I’ll fumble. I’ll repeat steps. I’ll act exactly like the kind of founder that devs love to complain about on Twitter.

Why?

Because I want to find out what the journey looks like when you're not technical-but you still want to build. I want to see how far AI tools can take someone who just has an idea. I want to capture every frustrating, confusing, hopeful, and hilarious moment of that ride.

I will try to put myself in a role of a non-technical startup founder, essentially a normal person who came with a great idea that they think it will be the next Uber.

Also, one note. I won't be posting screenshots of what I'm trying to build, and there's a reason for that. This is something that I wanted to build for a while, so I'll try to keep the concept a little bit under wraps at least for now. Of course, given what we see everywhere, you will probably see that released on th emarket tomorrow, maybe, or within two days as they claim.

Day 1: The Adventure Begins

00:00 - Starting

I started by registering at lovable, which took only a couple of minutes and was very easy. After that, I started working.

I explained the whole vision, a relatively complex product to manage contractors for our own use, listed out all the features and let it work.

00:05 - First look at my new SaaS

The tool generated something that, to be fair, didn’t look half bad. It had the usual SaaS-y dashboard look that startups love. Not how I would personally structure things, but it worked. So nothing to moan about.

I've spent a couple of minutes clicking around, looking at the screen layout and the information. Everything seemed alright or at least good enough for the first go, so I turned my attention to the prompt.

The prompt suggests that I either connect Supabase or read documentation. Since my rule is that I won't read any documentations, I clicked on connecting Supabase.

00:10 - Connecting to Supabase

I went ahead, it allowed me to connect to Supabase. I, of course, needed to create a new account and new organization, go through all the proving connection at the first try it didn't work, but it connected successfully at the second try. Maybe I was just moving too quickly, but either way Supabase connected.

I closed the separate window and went back to the chat and said "okay done". It suggested setting up authentication. I said, sure-but with a twist. I want users to log in, but not sign up themselves. The idea is that contractors will get an invite link. No public sign-ups.

00:13 - SQL? Really?

I was presented with a very long SQL script. If I really didn't know what's going on, that would scare the hell out of me. But pretending ignorance, I said, "Okay, whatever."

To that, I got an option to apply it or open it in an SQL editor. I said, "Just go ahead and apply it," and it did.

00:17 - Something happened… maybe

The database changes were applied. It told me authentication was set up, database integrated, next step ready. And then boom, login page appears, but with a nice fat "build unsuccessful" error in the chat.

I clicked it. Lots of info. No clue what it meant. Probably something to do with the database since I've seen SQL few times.

It asked if I wanted to create users or adjust "auth settings." What’s auth settings? I guess I’ll create a user.

00:21 - I have an user

I got redirected to Supabase’s user management page., created a user, closed the window and went back to the chat.

I then looked at auth settings out of curiosity. From the list of various sites I would guess it lets you sign in with different services. But when I clicked on Google, I got a pop-up asking for some API keys and such. Since I have no idea what it is and I don't have them, I just ignored it.

I have noticed however that "Allow new users to sign up" was still turned on. That’s not what I want, so I toggled it off and saved, went back to the chat and said: "Yeah, I did all that. What now?"

00:25 - The circle begins

I got an immediate response that it's trying to fix errors. I was mildly impressed that it actually detected an error. It detected what sort of error it is (something about types) and tried to fix it. After it finished, there was still a message built unsuccessfully, and the error messages were still exactly the same.

I decided to completely ignore it and said, "All right, what's next?"

It said: "Now you pay."

Since I couldn't do anything more, I decided to play around with it. I managed to log in with the account I have created before so that's good first step.

The dashboard still looks the same, so I clicked at the list of contracts and got a message "data is loading" but nothing happened. I've looked at a few other subpages and got the same, nothing is loading.

Guessing that this may be because I don't have any entries created. I decided to go ahead and create a new item. I got a nice looking pop-up with some basic validation. Not exactly what I wanted but good enough. My immediate thought was, that there will be a lot of interacting just around this one single pop-up. That is going to use a lot of prompts, so I will probably have to build prompts very, very big with a lot of information. Otherwise, I will run out of them before I even manage to get this one single pop-up the way I want it.

Anyway, I filled the form, clicked Save and got a message that a new item has been added. Went to the relevant list, still nothing loaded. Whatever I created in a day one, it just doesn't work. Looks like I'm stuck.

Conclusion

Ads I saw on Facebook claim that 20-30 minutes is enough to build the full application. I can straigh away say that a pile of BS and creative advertising. In my half an hour experience I've got an application that doesn't compile and doesn't work.

Well, I guess that's end of day one. I have used all five of my free daily credits and got an app that won’t build. I will use another five free credits tomorrow and try to fix it.

The adventure continues:

• Day 1: You just read it
• Day 2: a Still Nowhere, Still Stuck