Where it all started.
The first real project I ever built — six months into learning to code. A Heath-Carter somatotype calculator. The vision behind AfitPilot was already there: take measurements about a person, compute something specific, give them back a meaningful output.
A 2019 artefact, preserved.
In the spring of 2019 I had been learning to code for about six months. I wanted to build something real — not another to-do list tutorial. So I picked the thing I knew best from my sports-science background: the Heath-Carter somatotype.
Somatotyping is a method that takes nine anthropometric measurements — height, weight, three skinfolds, two bone breadths, two limb girths — and turns them into a three-number profile of body composition (endomorphy, mesomorphy, ectomorphy). Carter and Heath published the modern formulation in 1990. It's still used in talent identification and sports research today.
I wrote the maths in vanilla JavaScript, hand-drew the somatograph in HTML canvas, found the academic references in my old university PDFs, and shipped it on GitHub Pages with a Bootstrap 4 wrapper. Looking at it now I can see every shortcut and every awkwardness of a first-time builder. I kept all of it.
Because the thing that matters isn't the code quality. It's the shape of the idea. Six years before AfitPilot existed as a company, I was already trying to do the same thing I'm doing now: take a set of variables about a person, run them through a system, generate a personalised output. My grandfather did it for heart disease in 1971. I did it for body composition in 2019. I'm doing it for athletic performance today. Same intellectual architecture, three generations apart.
Try it — exactly as it was in 2019.
This is the original code, hosted unchanged on AfitPilot. Bootstrap 4. jQuery. Vanilla JavaScript. A canvas element. The only thing removed is the Amazon affiliate panel for measurement equipment — those tracking codes are stale and no longer relevant.
Open the standalone 2019 pageSource, demo, and the original Heath-Carter manual.
The same idea, six years later.
In 2019, the calculator took nine numbers about a person and returned three. Useful, but limited — a snapshot of a body, not a plan for it.
AfitPilot, today, takes a much richer set of inputs — training history, movement capacity, daily readiness, fatigue markers, sport demands, recovery state — and returns a programme that adapts as those inputs change. The somatotype tool was a static reading. AfitPilot is a feedback loop.
But the underlying belief is identical: a system built carefully, from real measurements about a real person, beats a generic template every time. I didn't have the words for that in 2019. I just had a hunch and nine measurements.