December 2023. I decided to make my own game: one I would want to play myself and that, maybe, other people would want to play too.
I'm a software engineer, so I knew the technical path. I researched engines and settled on Godot: open source, and no fees on the game. I devoured YouTube tutorials and the official documentation, and as soon as I worked up the courage, I started building the first mechanics.
The inspiration came from way back. I spent countless hours on Warcraft 3 custom maps, Dota and an endless stream of tower defenses, and one of them, "Hard TD", became my first reference. The project was born under the name "Gread TD".
The first commits landed in January 2024:
And this is what the game looked like:
I'm also a roguelike fan, and I wanted that flavor in the game. Design was never my strong suit: I tried teaming up with two friends who knew more about it, but they didn't have the time or the drive to stay on the project. So I kept using free assets from the internet, just so I could see what I was building.
Card picks arrived between waves. To raise the tension, every card came with a chance of losing a tower. That idea later became a worm that bursts out of the ground and swallows the whole tower.
The first card selection layout:
The first skill tree:
I'm a huge Factorio fan, and I wanted to introduce people to the genre without scaring anyone off: just a few pieces, easy to understand, but ones that make a real difference in a match. That's how the Collector tower was born. At first it only gathered insect remains to feed the factory; later it learned to clear destroyed towers and rocks off the map.
The first factory:
From the start, I dreamed of a platform where anyone could build the tower defense they wanted. The original idea was a DLC with an editor; in the end, I decided to open everything up to mods on launch day.
At some point I realized the game could be more than 2D. The mechanics were already solid, and throughout 2025 I tried several times to bring AI into development, but the technology wasn't mature yet: it was faster to do everything myself.
That changed in early 2026, when LLMs started writing genuinely good code. I began using them for balancing, a hard task I'd been banging my head against for years without ever finishing. I soon realized I could get help with the visuals too, and that's when the full transformation happened.
From this 3D:
To the first version of the cockpit:
To this:
The cockpit card exists because of my son. He was always pushing me to go further, and one day he asked to control a tower. I said no: that wasn't the kind of game I wanted to make. But the idea kept nagging at me, and I ended up adding it as a temporary card.
After that, I decided to see how far Godot could go. Why not switch to a realistic palette, then refine and optimize everything, until it became one of the best-looking games ever made with this engine?
Torreta launches on October 9. See you in the pit!