My wife and I own a Tonal. We've used it religiously for six years. I've more than tripled my strength score on it. We love the machine.
But we both wanted something the machine — and the rest of the fitness software market — couldn't quite give us. We wanted training that was actually built around our own data.
The data was rich. The action was missing.
Every workout I did on the Tonal was logged. Every set, every rep, every weight, every progression. Years of training history sitting in an app I could open any time. I knew the data was valuable. I just couldn't do anything with it.
So I went deep. I extracted my data through the API. I analyzed it spreadsheet by spreadsheet. I learned a lot — which lifts I had stalled on, which muscle groups I'd been quietly ignoring, where my volume was light, where my progression curves had flattened. It was real insight.
And then I hit the wall: now what?
Knowing what to fix is one thing. Translating that knowledge into a workout — a real one, on the Tonal, ready to go when I walk up to the machine — turned out to be a separate, much harder problem. The Tonal's built-in workouts are great, but they aren't designed around my specific data. To get something custom, I had to build it manually. Each exercise. Each set. Each rep target. Each weight. For every workout in a cycle.
It was painful enough that I mostly didn't do it.
Two more stories that pushed us over the edge.
The trainer translation problem. I worked with a personal trainer for a stretch. Great trainer. He'd send me workouts every week, written for whatever equipment a normal gym has — barbells, dumbbells, cable stacks. The Tonal can replicate most of that, but it requires translation: which Tonal movement maps to a barbell row? Which is closest to a cable fly? How do I dial in the right weight given Tonal's variable resistance? I spent more time translating his programs into Tonal-executable form than actually training. The coach was great at his job. The infrastructure between his job and my machine didn't exist.
The broken-foot scenario. Last year I broke my foot. I still wanted to train — most lifters do, when they can. The Tonal didn't have a program designed for someone with a non-weight-bearing lower body. I asked an AI for a custom program, and it gave me a great one: seated, upper-body-focused, structured, well-progressed.
Beautiful program. Useless on the Tonal.
To actually use it, I had to build every workout by hand on the machine. It took me days. And after I'd done all that work, I realized I'd missed something obvious: I hadn't grouped exercises by accessory. Every workout had me swapping between handles, ankle straps, the bar, and the bench multiple times mid-session. The program was biologically sound. It was operationally a mess.
That was the moment the idea sharpened. The data is sitting there. The science is well-understood. AI can custom-build programs for any constraint a human has. The piece that was missing was the layer that turns all of that into something executable on your specific Tonal — automatically, every cycle, without you having to translate a single thing.
From a tool for us to a product for everyone.
We built the first version just for the two of us. Quietly. As a side project. I extracted our data from our Tonal, built the analysis, generated the programs we wanted.
Then we showed it to friends — other Tonal owners. Same reaction every time: I want this. People who already loved their Tonals. People who tracked their sleep and their steps and their workouts but couldn't connect those data points into a real program. People who'd worked with coaches and felt the same translation friction. People who, like us, had walked up to the Tonal hundreds of times and “picked something.”
The pattern was clear enough that we decided to build it properly. Not as a script for two users. As a real product, for anyone who has a Tonal and wants their data to actually do something.
What we believe, simply.
We believe the data already exists. We believe the science already exists. We believe the AI tools to fit one to the other now exist. The piece that's been missing is the layer that puts them together, on your specific machine, automatically, every cycle.
So we built it.
If you have a Tonal and you've ever felt the same frustration — that your data is rich and your action is poor — we built Trophic Labs for you, too.