I had this thought recently about whether I could actually take an idea, something half scribbled in my notes, and turn it into a working prototype in the space of just a couple of days. Not something polished or ready to ship, but something that works well enough to prove the concept and show me that it has legs.

The problem I wanted to solve was one that I’d been experiencing myself for a while, Reddit is incredibly powerful when it comes to feedback, I’ve written about this recently, but it’s also famously brutal. While the value is definitely there, the way it’s delivered can be overwhelming, distracting and honestly a bit demoralising. So this was my attempt to take all of that value and reframe it into something more useful, something that kept the insight but lost the hostility.

So I thought, why not use this as a way to test some of the new tools that have been constantly coming up on my feed, things like Lovable, Cursor and MCPs. I’ve used Cursor for a few years now, last time I used it though, AI coding assistants felt clunky and unreliable. As soon as a project got too big, it would get confused. Things seem to be moving fast in recent months, I knew I needed to get myself up to speed, especially now that tools like Lovable let you generate and iterate code, spin up a backend, and push everything straight to GitHub almost instantly.

The first build

I started properly, with a PRD, a full project requirements document laying out exactly what the tool needed to do and how it would work. Next step was to drop that into Lovable, which surprisingly came back with a first version of the homepage along with a Supabase backend already wired up. The part that really got me though was that Lovable connects straight to GitHub, which meant I could instantly take that codebase into Cursor and start shaping it, feature by feature, without losing momentum.

By the end of that first day I already had a functional tool, and it genuinely worked, but it looked and felt like the most generic version of itself. Very template-driven, the standard shadcn UI, and while it was functional it didn’t carry any of the personality or design intent that I wanted.

Where MCPs changed the game

This is the point where the rise of MCPs really came alive for me, Figma released their MCP a little while ago, and I thought why not take advantage of it. So I put together a proper branded UI in Figma that reflected how I actually wanted the product to feel, then connected that design back into Cursor and asked it to rebuild the front end based on my Figma file.

It wasn’t perfect, there was a fair amount of back and forth and a few features got lost along the way that had to be added back in, but the results were undeniable. The fact that I could move between design and working code almost seamlessly was incredible to see in action, and it really did feel like a different way of working, one where the distance between idea and execution has been compressed almost beyond recognition.

What it shows

In the end it took me maybe two or three days from start to finish. I went from having the idea, writing the PRD, generating the backend and the homepage, to building out the features. Then on a whim, plugging in Figma and seeing it redesign the whole thing based on my flat designs, and to having a working prototype that I could actually show to people.

Not long ago that whole process would have been weeks of work, maybe even months if you factor in the handoffs and the typical loops between design and development. Now it can be an evening project, and that shift says a lot about how prototyping and building digital products is changing. Tools like Lovable, Cursor, Supabase and these new MCPs are not gimmicks or nice-to-haves. They are collapsing the space between thought and thing.

Of course this doesn’t replace design craft or proper engineering expertise, but it does fundamentally change the rhythm of how ideas can be tested, and what used to be a huge lift is now small enough that you can go from “I wonder if this would work” to “here’s a working version that you can try” in the space of a weekend.

It feels like we’ve stepped into a completely different universe, and I think we’re only just starting to see what’s possible.

If you’re curious, you can see the video of what I built below. If you want access drop me an email and if I get enough, I’ll do a full release!