Building ScoutHalo.

In October 2025, I wrote down an idea: an app for finding specific places for different kinds of creators. It was one of those random ideas you capture before you forget it. Then I left it there.

That summer, I was working at a friend's pub. When the season ended, I wasn't entirely sure what I was going to do next.

I started spending every day working with AI and learning what I could build with it. A few weeks later, I went back to that note and decided to give the idea a shot.

I started small: a landing page, some early experiments, and eventually the first version of what would become ScoutHalo.

The idea eventually became something I wanted to test.

I wasn't a software engineer, but AI gave me a way to start building it myself. The first version was called ScoutNYC. The plan was simple: start with one city, build a useful collection of shoot locations, then expand to Los Angeles, Miami, and beyond.

I built the first website and started adding locations manually. Each one needed the information I thought a creator would need to actually use it — where to shoot, how to frame it, when to go, and what could go wrong.

It worked well enough to prove the concept to myself.

But it also exposed the problem with the way I was building it: I was the database.

Every new location required me to find it, research it, structure the information, and upload it myself. Getting from a handful of locations in New York to thousands around the world wasn't going to happen that way.

At the time I also needed to make money and had other projects competing for my attention, so I put it aside.

Eventually, I found a better way to build it.

I came across a different approach to building with AI that made the part I'd thought was impossible suddenly look solvable.

The problem that had made the first version impossible to scale started to look solvable. I realized I could build the product differently, and this time I decided to go all in.

I spent most of that winter in Uruguay in one small room, with a bed next to my desk and a computer. For about three months, most days started when I woke up and ended when I went to sleep.

I was learning as I built. AI let me move at a speed that wouldn't have been possible for me otherwise. Some days I had as many as four Claude Code agents working in parallel on different parts of the product.

I wasn't working a job during that period. My family made it possible for me to take that risk. They supported me emotionally and financially, helping cover my basic living expenses and the tools I needed to keep building.

That support gave me something incredibly valuable: time. I used it to turn the rough idea I'd written in Notes into the first real version of ScoutHalo.

What started as a sentence in my Notes became ScoutNYC, and eventually ScoutHalo.

The product today is very different from that first version. ScoutHalo is a location intelligence platform for production teams, designed to help them discover, evaluate, and prepare locations before a shoot.

I'm now working on the next part: getting it into the hands of production teams, learning from how they use it, and building the company around it.

This is still the beginning.