Zero to Production App in 7 hours with Lovable

Last week, I finally decided to give Lovable.dev a go on a little project that’s been on my list for years. I’ve always wanted a Reddit-style user feedback system for my projects, something that lets people vote on ideas and comments. It wasn’t the smoothest process, but I wanted to do it entirely without writing a single line of code. My goal was to prove that “vibe coding” (fully coding using only AI and no technical knowledge) is actually feasible.

Not only is it real, but it’s significantly faster than building from scratch. To emphasize this, I let Lovable choose all the frameworks and the entire stack.

Including buying a domain name, tinyupvote.com, the site cost me a total of $35. It’s pretty basic, but it’s functional. 

And it all came together in just 7 hours, from zero to launch.

Once I got it together, I couldn't resist creating a few roadmaps. (Hopefully I'm not the only one who enjoys this kind of system.)

I created a Tinyupvote roadmap for the next features I’m considering for Tinyupvote itself (meta, I know):
Next Tinyupvote Features

I also created one to figure out what I should write next:
Next Writing Topics

And finally, there’s one for the features I should add to Wordara:
Wordara Features

Please feel free to vote or add comments if you have thoughts! You can also create your own Tinyupvote.

I love how Reddit comments work and how users can riff off each other recursively. At the same time, ideas with merit (or at least interesting ones) rise to the top. I wanted that same dynamic. I searched for a plug-in or embed to replicate it, but everything I found was either too complicated, too janky, or too expensive for the stage my projects are in right now.

One tough decision was requiring registration. I’m still on the fence, but I think any voting system would be meaningless without tracking genuine human votes. Another interesting idea is allowing the author of a Tinyupvote (on an opt-in basis) to access people’s emails. This could be important because certain people (for example, actual customers) might be worth paying special attention to.

The Stack
In true vibecoding fashion, I actually didn't know and had to look it up for this article. Fortunately it is self documenting in the readme. Or at least I think it is... hopefully these aren't hallucinated:

It also uses Supabase for storage and authentication, which is a concerning omission from the auto-readme.

Overall Lovable is pretty lovable. A number of people have told me they were frustrated with the results, and I agree. But it's important to remember coding was frustrating without AI too. Getting stuck is a natural part of the process. However the frustration to progress ratio is very good. One thing I really liked was how each change was a commit and it synced to Github.

When you hit a wall
It's painful because the promise is there. You get something that appears 80% complete in 15 minutes, but getting it to a deployable state takes considerably more work. I would highly recommend pairing Lovable with Cursor, because Lovable can (and will) get stuck. When it gets stuck on an issue, pulling the project locally and then asking a different model what was wrong worked pretty much every time. Once Cursor had a fix, I could push the change to Github, Lovable would auto-sync it and usually Lovable would get unstuck.

The annoying rewrite issue
One infuriating thing that every AI coding solution I've ever tried does is fixing one issue and deleting another along the way. The solution to this is testing every major change and rolling back if there are issues. Ideally you can have it write tests (just like in normal software development) that will confirm everything is working as expected, but if you are just messing around testing by hand can do the job too.

Maybe it's not fair to say a non-technical person can do all of this in 7 hours, but I think with a bit of a guide or even some research it's totally possible. It's also arguably unfair to say that I built it for $35 because Lovable is a recurring subscription and I'm not factoring in Cursor which I already pay for. That being said I think I will continue using Lovable. I have a bunch of ideas and it helps me ship MVPs fast.