The origin of App Composer

I kept finding ways to make things without building every piece myself.

I didn't start with a grand plan to build a software platform. I had things I wanted to make, and a habit of finding the right tools before writing my own.

That habit followed me through the Navy, IT, music, online stores, DevOps and eventually AI agents. App Composer is my attempt to make it useful to someone else.

01 / The long way round

Music got me building for people.

I served in the Royal Australian Navy from 2004 to 2010, then spent the better part of the next decade working with desktops, networks and systems. I moved into DevOps in 2019.

Around the same time, the music I was involved in started reaching more people. That meant websites, releases, merchandise and stores needed looking after. I learnt WordPress and WooCommerce, built stores for Shake Down Records and Grind Mode Cypher, and helped with the social side too.

Somehow, between music and keeping the technology running, I ended up making software.

02 / The method

My first trick was remembering what was good.

When I was working with WordPress, my approach was a pretty practical three-tier system:

  1. 01

    Reuse what works

    Find a good plugin or component. If it was something I'd need again, a lifetime licence often made sense.

  2. 02

    Pay for the specialist part

    For an occasional need or a feature I couldn't reasonably build, use the right paid tool rather than pretending I could do everything.

  3. 03

    Build what's missing

    Write the custom part when the available pieces couldn't do the job. Eventually that meant learning to code properly.

I did much the same thing with GitHub repositories. Sometimes knowing which project already solved a problem made me look considerably cleverer than I was.

But finding the right thing, understanding whether it fits and knowing what still needs building is useful work.

03 / Bigger systems, same principle

The pieces got bigger.

As my work moved deeper into DevOps, I worked with Ruby on Rails, AWS and infrastructure migrations, including moving applications away from Heroku. The problems changed, but the instinct didn't: reuse the good parts, connect the systems carefully and automate the work you shouldn't have to repeat.

AI agents changed how much I could attempt. I could describe an outcome, show examples, ask agents to investigate and implement it, then review the result. The important parts didn't disappear: judgement, taste, permissions, testing and knowing whether something actually works.

04 / Why App Composer exists

What if everyone had that workshop?

I wanted the things I'd spent years figuring out to be available to anyone with an idea: discover useful building blocks, explain what you want, bring your own agent, make the missing pieces and end up with something you can actually keep.

Not just an installer for software that already exists. A place to start with something that doesn't exist yet — an app, a game, a creative project or a workflow — and work towards building, running and improving it.

That's the direction of App Composer. The current product has working tools and examples alongside features still being built; it isn't a promise that every idea can already be made and deployed in one click.

Find what works. Make what doesn't exist. Keep what you build.

Explore the working toolsSee a real buildStart with an agent