After I decided to try building a world around my album SCENES with the help of a small team which included a designer and a developer , the next question was: how far could I take it myself?
No team, no developer. Just curiosity, YouTube, and some AI code tools.
My first test run for this was to try and build a game as a simple way for people to interact with my song, which you can play here and read more about the ‘why’ here if you missed it.
I almost forgot to post this post, but I’ve had a few people asking what tools I specifically used and how I made the game, so here is a quick break down below. I wanted to share just to show why I believe anyone can create ideas that are outside of the normal expectations around music releases. In the coming weeks I’ve invited some other artist to share what they are trying too, so debunking that people felt building the game was beyond them felt a good place to start.
The game worked really well in the end, a very small number of people had a go at it, but those who did got to win free merch, get a free download of the song before release day and I got a little closer to connecting with those that cared about my music more… which has become so important for all creators.
1. Finding the starting point
I began by opening a few platforms, first Claude, then Vercel VO, asking them to make “a simple game.” - These are both LLM platforms like Chat GPT that write code.
The little games they created looked fine, but what I found hard was how to customise it because it was literally made from code and pixels, so while killing time I turned to YouTube tutorials to try and get a better idea of what to do next.
I found a video called “Build a Simple Shooter in Cursor” which stood out as I could see that the user was customising the game look more than i had been able to do. The You-tuber used something called Vite, which I’d never heard of before. I took the same approach that I often take with anything new, and just followed each step he did pausing the video as I went along when I needed too.
I will keeping reminding you that I have no idea what I’m doing. It’s important, because if you haven’t tried playing with no-code tools I really think you should have a go. I even remind what ever platform I’m using of this all the time, and you may have to too :)
For example i had not idea what Vite is or does so i found out by literally asking Cursor: “What is vite and why am i using it?”. The answer: “Vite (pronounced veet) is a lightweight web-build tool, basically a fast engine for running JavaScript projects in the browser”. The useful bit I realised was it let me preview the game instantly and get a feel for what I was making in real time. That immediacy was addictive, and much more natural to me, just like making a beat.
2. Shaping the gameplay
The YouTube video I was referencing had a little player shooting blocks, but I wanted it to feel like my music: fast-moving, rhythmic, constantly in motion.
So I decided to go more in the direction of a runner game. The player keeps moving forward, dodging obstacles and collecting things. I just said to the platform something simple like “I want to make it like the old Mario Brothers not a shooter”..
One of the advantages of experimenting with LLM’s like Chat GPT and Claude is they can help you write better prompts, which in turn give you a more solid platform to build from. In this case I wrote down the tools the coder in the YouTube tutorial suggested, and asked Claude chat write me a ‘better’ prompt to build the game, that I could then copy and paste into Cursor. This can be helpful, but also sometimes all these tools work better if you keep it simple and just say what you want the end product to be and the experience you want the user to have. Try both, see what works best:
3. Creating the player character
I needed a lead character, so I jumped into ChatGPT and Gemini to design a small pixel-art DJ, the avatar of my world.
They helped me rough out colours and shapes, but when I asked for a running animation, the models struggled.
I took the sprite into Canva, redrew the legs in three different running poses, exported each frame, and removed the backgrounds manually.
That became my three-step run cycle.
4. Adding things to collect
Next I designed nostalgia items like MP3 players, vinyl, and old photos, representing the memories we’re all trying to “get back” to.
These became the collectibles that give you points.
Then came the baddies: phones, adverts, and random digital clutter.
At one point I even added simple blocks to jump on, creating small moments of pause between chaos.
5. Building bosses
After a few play tests I wanted more of a story.
I decided the game should have bosses, larger enemies representing the “press-platform villains” that fill our feeds.
When ChatGPT refused to generate certain characters ( insert eye roll emoji) , I switched to Gemini, which helped me create the pixel versions.
I then added a point system so that when a boss appeared, it triggered a score multiplier and sound cue.
6. Making it sound like the song
I looped the full track in the background, but the sound effects didn’t feel right.
So instead of using stock bleeps, I went back to the original session for I Wanna Get Back and exported short synth bursts, kicks, and vocal chops.
Those became the in-game sounds, meaning every effect literally came from the song.
It made the music and the game feel like they were talking to each other.
7. Adding structure: intro and ending
I wanted it to feel like a complete piece, not just a test level.
I prompted ChatGPT to generate a custom intro screen with the title and instructions, then an end screen that appears when you finish.
Now it felt like an actual release, a mini-experience around the single.
8. The leaderboard and rewards
Next came the feature that turned it from a toy into a community experiment: a leaderboard.
I built it with Supabase, which stores player scores and emails.
This allowed me to send everyone who played a download link for the song as a small thank-you for taking part, and to reward the top scorer each day.
9. Hosting it publicly
To share it, I needed a simple web page.
I used Vercel to host the build and attach a domain so anyone could play directly in a browser.
That step made it real. I could finally send out a link rather than a folder of files.
10. Automating the connection
During SCENES, one of the hardest parts was keeping up with fan submissions, creating art, emailing people, and documenting everything while also promoting new songs.
This time I wanted the game itself to build connections for me.
Using Make.com and Brevo, I set up automation so that when someone submits a score:
their data goes to Supabase
an email triggers automatically with the song download and streaming links
I get notified if they’ve hit the top of the leaderboard
Now the system keeps engaging people even when I’m not online.
11. Designing a reason to keep talking about it
I wanted the game to feel alive day to day.
So I created a sticker sheet of all the pixel characters and nostalgia items, unique to the game’s world.
At the end of each day I check the leaderboard, announce the winner on socials, and mail them the sticker pack.
It’s small, but it creates a rhythm of play, reward, and conversation.
12. The bigger picture
Each step started as curiosity and turned into design.
By the end I have a functioning mini-world that connects to my song both sonically and in concept.. hopefully!
I want the process to be a reminder to me and others that we don’t need big budgets or external teams to make interactive art anymore.
For me, this is a good example of what digital world-building can look like:
small, handmade systems that let our work connect with people automatically, playfully, and meaningfully.
🧰 Tools Used
Cursor + AmpCode – my AI-assisted code editor. AmpCode handled most of the JavaScript scaffolding and debugging while I described what I wanted in plain language.
Vite – a lightweight web-build tool that lets you preview code instantly in the browser. It kept the feedback loop fast and fun.
ChatGPT / Gemini / Claude – used interchangeably to generate and refine pixel-art sprites, screens, and short code fragments whenever I got stuck.
Canva – where I redrew the DJ’s three-frame run cycle and cleaned up sprites before exporting them with transparent backgrounds.
Supabase – database and auth layer for storing player scores and emails, powering the daily leaderboard.
Make.com – automation glue connecting Supabase to my email platform.
Brevo – sends the automated “thank-you” email with download and streaming links.
Vercel – hosts the finished game and lets me attach a custom domain for public sharing.
Ableton Live – where I pulled stems and short sound effects from I Wanna Get Back to make the game sonically part of the track.
Sticker Mule / Canva Print – for producing the daily prize, a sticker sheet of all the game’s characters and nostalgic items.
💿 Play the Game + Join the Experiment
If you missed the first post, I Wanna Get Back is my new single, and this game is the world I built around it.
It’s a retro runner where you collect memories, dodge the noise, and feel the music interact with the visuals in real time.
🏃♂️ Play the game here
This whole project is an experiment in how artists can build small worlds around their music, worlds that connect, reward, and play back.
Thanks for being part of mine!
My next project is kinda like a SCENES part 2, and I want to invite you to help me test it, work on it and take part in the album. if you are interested you can sign up to join the updates channel and the exclusive chat group HERE
Off For New Shores! 🤘🏻💛