Adventures in Vibe Coding
From experimentation born out of necessity, here's my journey (so far) into the mind blowing - and sometimes mind-boggling - world of Vibe Coding!
I’m sure many of you who’ve worked in a corporate environment have had this thought: “If only we had a tool that could just do this.” Then reality sets in - suggesting something new means months of bureaucracy and, nine times out of ten, nothing ever happens. Or worse: by the time a solution is approved, the original problem has already changed.
At the beginning of this year, I’d paused work on Imbr and knew I needed to prioritise other ways of generating income. That meant being flexible with how I worked, and I definitely didn’t have the budget to cover the hefty price tags of most SaaS platforms pitched as “one-stop solutions” for the music business (spoiler: they rarely are). My options were to go without, overspend on features I didn’t need, or find a manual workaround that would take ten times longer than it should.
What struck me was just how fast AI has moved. The landscape today is unrecognisable compared to even a year ago - let alone three and a half years ago, when I left Warner Chappell. I’d dabbled with no-code tools before; they’re great for quick prototypes or scrappy first versions. Platforms like Bubble and Lovable are perfect for getting your feet wet, but they’re not built for deep functionality or polished design. That’s not a criticism - they serve their purpose.
The real turning point came when I read a VC’s prediction that it was only a matter of time before a non-technical solo founder built the first Unicorn (**a start-up worth more than a Billion dollars). I didn’t think, “I could be that person.” My thought was more, “What would it actually take for that to happen?” And that lit a fire of intrigue.
That’s how I found myself diving headfirst into vibe coding. At first, I thought it would just be a bit of fun (and it is - a lot of fun, actually). But recently, when I faced a five-hour manual task, I decided to see if I could build my way out of it. Forty-five minutes later, I had a tool that automated the entire thing. What used to take me five hours now takes under twenty minutes.
That’s when it clicked: vibe coding isn’t just a clever hack. It’s an entirely new way of working.
So what is Vibe Coding?
For me, vibe coding sits somewhere between no-code and full-stack development. No-code tools are great for prototypes or scrappy MVPs - I’ve used them, and they’re brilliant for getting something in front of you quickly. But they’re also restrictive. The moment you want more flexibility, more polish, or a specific workflow, you hit a wall.
That’s where vibe coding feels different. You’re not dragging boxes around a canvas or locked into someone else’s templates. You’re working with AI in a way that gives you the freedom of code, but without needing years of experience as a developer.
Some days that means I’m sat in VS Code, going back and forth with an LLM, writing line after line until something finally works. Other days, I’ll use something like Cursor, which supercharges the process because it understands the entire codebase at once. (And yes - it’s the first time I’ve ever seriously considered paying $200 a month for a piece of software.)
What I like most is that vibe coding isn’t about building flawless products. It’s about building the exact tool you need, when you need it - even if it’s scrappy, even if it’s not “scalable.” And that’s the point. It solves the problem in front of you, right now, so you can get back to the work that matters.
First Attempt: Matchmaking Tool for Songwriters
The very first thing I built was a songwriter matchmaking tool. The logic was simple enough: connect writers in my database based on different attributes, to help set up sessions or camps with natural fits.
Locally, it worked beautifully. I was convinced I’d cracked it. Then I tried to move it online… and everything collapsed. None of the logic held up, and two weeks of debugging later, I admitted defeat.
But I came out of it better off - I learned the principles of how to debug, how to give AI clearer instructions, and even how to make an interface look clean instead of like a side project held together with duct tape. I’ll rebuild it from scratch eventually, but the lessons from that first adventure still shape how I build now.
A More Useful Win
One of the pain points I ran into early on was generating CWR exports for my publishing admin company. For those not familiar, CWR is the standard format used to report works to collection societies. Every SaaS platform offers this, but most hide it behind a paywall and call it an ‘add-on.
That didn’t sit right with me. I’d already built my own backend admin platform - why should I pay extra just to export data in a standard format, using a tool that duplicated what I already had? No thanks!
So, I decided to see if I could build it myself as an add on to my admin platform. It wasn’t exactly glamorous - there’s nothing particularly exciting about data exports - but with a bit of back-and-forth in vibe coding sessions, I had a tool up and running that automated the whole process.
Now, instead of paying a monthly premium, I’ve got a system that works exactly the way I need it to - and it costs me nothing beyond the AI subscription I was already paying for.
Going Bigger
Once I’d seen what was possible with smaller tools, I wanted to push myself with something more ambitious. Up to that point, vibe coding had been about automating admin and building utilities. But I also wanted to centralise all my business verticals in a more streamlined way - so I asked myself: what if I rebuilt my entire website?
So, I did.
From scratch, I built a site that wasn’t just a static front page. It has funnels, sign-ups, merchant tools, upsells, and a content management system all integrated. You can see the different facets of my businesses, buy products, join my publishing company - and every part of the workflow connects to the next.
This wasn’t about being clever for the sake of it. It was about creating a system that could support the business end-to-end, without the patchwork of third-party services I’d otherwise have needed.
It took some trial and error, but the result works. It’s functional, it’s mine, and it was built faster and cheaper than I could have managed through traditional routes.
And honestly, the fact that I - someone who is 100% not a developer - rebuilt an entire business website from scratch still blows my mind a little.
My Growing Toolkit
Since then, vibe coding has become my default approach. Here are a few of the tools I’ve built along the way:
- A publishing admin platform
- CWR export tools for reporting to societies
- A data compiler, exporter, and reformatter
- A full website with funnels, sign-ups, merchant tools, and upsells
- A Spotify data analyser
- A songwriter project management tool (currently moving from mock-ups into build)
- A writer portal for my publishing company (in testing)
And outside of music industry focused projects, there’s:
- A Japanese-language chatbot – and anyone who has tried to learn Japanese understands the complexities of reading the language, so this fixes that and allows you to have full conversations with it in a natural way.
- A career coach chatbot integrated into WhatsApp, based on my book.
Some of these tools save me money. Others save me hours every week. All of them give me independence. And this is only the start - my backlog of ideas keeps growing, and the creative energy this work unlocks has made even the most mundane tasks feel exciting again.
Why This Matters
Inside corporates, getting a new tool approved is painful. I once watched a project management tool take two years to roll out - mostly because the wrong person didn’t like the idea. By the time it finally launched, the momentum (and the need) had long gone.
Independents don’t have that problem. If I see a gap, I can build something to fix it - quickly, cheaply, and in a way that actually fits my workflow.
That’s the real advantage. Vibe coding doesn’t create flawless products; it creates practical ones. It lets you move faster, capture value sooner, and skip the bureaucracy. And it works just as well for creators who simply need to be more efficient with limited time and endless to-do lists.
For me, vibe coding started as an experiment. A survival tactic to avoid spending money I didn’t have on tools I didn’t need. But the more I’ve leaned into it, the more I see it as a genuine shift in how anyone can work.
I’m considering putting my experiences into an eBook - not a technical manual, but a practical guide for anyone curious about building their own tools. If that’s something you’d find useful, let me know.
Because if vibe coding has taught me anything, it’s this: you don’t need to wait for someone else to solve your problems. You can just build your own solutions. Sometimes all you need is a nudge to get started.