character design for smokee armour
companion/card for the in-app game
minotaur that you will battle
how the character looks below level 5

The entire project was built by AI. The iamges, code, videos - everything. It's still in development.

HeroRise

What inspired me

I originally started building HeroRise for myself.

I'm a dad who likes gaming, gamification and a really good storyline. I also like the idea of improving yourself and seeing that progress in a way that feels real.

When I looked around, I found habit trackers and apps that gave you points for completing tasks. I also found AI coaching apps and games with stories. What I couldn't find was something that brought all of that together.

I wanted the things you do in real life to build up a hero inside the game. You complete missions, earn XP and gold, improve your stats and unlock more of the story. Eventually, the AI coach will help people work through whatever is stopping them and give them personalised missions based on what they need.

The AI coach hasn't been added yet. The app has been built with it in mind though, and some of the things it will need are already there.

I knew from the start that HeroRise could work for other people, but I still made it around what I wanted. After doing more research, I found there weren't many direct competitors. I think there are other dads like me who want to improve themselves but would enjoy it a lot more if it felt like playing a game with a story they actually cared about.

How I built it

I started building HeroRise in Firebase Studio. I had never built an app like this before, so I was learning while I was making it.

Firebase Studio worked well when the app was smaller. Once it got bigger, it started making too many mistakes. Something would get fixed and then something else would break. It got to the point where I needed more control over what was happening.

I moved the project into VS Code and started using OpenAI to help me build it. I tried Claude for a while as well, but I kept coming back to OpenAI.

When Codex became available, I started using that instead of working mainly through VS Code. I'm not sure exactly when I made the switch, but Codex became a big part of how I built HeroRise from that point.

A lot of the process has been me asking for something, testing it and then working out why it didn't quite do what I wanted. Sometimes it worked straight away. Other times I had to go back over the same feature several times before it was right.

I've also built a few smaller projects along the way. Most of them were tools for my work, productivity or focus. HeroRise is still the main project and the one I keep coming back to.

What I learned

When I started, I didn't really understand how all the files fitted together. I could ask AI to make something, but I couldn't always tell what it had changed or why.

I have a much better idea now. I can look at the structure of the project and roughly understand how Codex is putting things together. I'm still not an expert coder, but I'm also not blindly accepting whatever it gives me anymore.

Prompting has been a huge part of what I've learned. I found out quite quickly that asking for something in a simple way doesn't always get a simple result. I had to get better at giving context, explaining exactly what I wanted and making it clear what shouldn't be changed.

I also had to learn about APIs, Node servers, hosting, storage and deployment. Before this, I didn't know much about any of that. Building the app was only one part of it. I still had to work out how to get it online and make everything work once it was there.

The challenges I faced

To be completely honest, working with AI can be frustrating.

Sometimes I ask it to change one thing and it changes something else as well. Sometimes it misunderstands a clear instruction and does something that makes no sense. Then I have to work out whether the problem came from my prompt, the existing code or what the AI decided to do.

As HeroRise got bigger, that became harder. It has quests, player stats, rewards, combat, cards and a story that keeps progressing. All of those parts affect other parts of the app. A small change can cause a problem somewhere I wasn't expecting.

Learning the technical side while building the actual project was probably the biggest challenge. I wasn't only learning how to code. I was learning how a full app fits together and how to take it from an idea on my screen to something another person can actually use.

I'm still learning and HeroRise is still being built. It is already much further along than I thought I could get when I first started, especially considering I had never done anything like this before.## Inspiration