SomataQuest started because I was tired of taking screenshots.
Every day, I was taking pictures of two different apps—both of which I paid for—and feeding those screenshots into ChatGPT to help decide how I should train. I was tracking strength workouts in a third paid app. Between them, I had plenty of data, but no single place that brought it together around the question I actually cared about:
Given my recovery, recent training, sleep, and goals, what should I do today?
There had to be a better way. I went looking for it, but I could not find an app that worked the way I wanted. Most gave me a headline score or a generic assurance that I was probably fine to train. That was not enough.
Sometimes I work late and do not get enough sleep. Sometimes one recovery signal looks good while another does not. In those situations, I do not just want an app to say, “Meh, you can train.” I want to understand what the available data says, how the different signals fit together, and whether a hard session is likely to be productive.
So I started building SomataQuest.
In my career as a physical chemist at a national laboratory, I care deeply about using the best available data to make defensible decisions. I brought that same mindset to SomataQuest: combine the relevant information, make the reasoning transparent, and ground the approach in current research.
SomataQuest brings my sleep, recovery, activity, and training data together in one place. It considers the broader pattern rather than relying on a single headline number. More importantly, it helps me understand why it is recommending a particular kind of training—not merely whether I have permission to exercise.
The app is also configurable because there is no universally correct way to train. If my goals change, I can reflect that. If someone follows a different training style, that can be reflected too. SomataQuest does not assume that a population average defines what is normal for every individual. Wherever possible, it evaluates changes relative to your own baseline, because the useful question is not simply how you compare with everyone else. It is how you are doing relative to yourself and what you are trying to accomplish.
I also wanted the data to remain useful beyond the app. SomataQuest can organize it into clear, entirely text-based summaries that work with whichever large language model you choose. If you prefer ChatGPT, you can use that. If you would rather use a local model and keep the entire process private, that works too. Your data should not be trapped inside someone else’s interface.
I now use one app instead of three. The information I care about is at my fingertips, and I feel much better informed when deciding whether to lift, train easily, or recover.
I originally built SomataQuest to solve my own problem. I am sharing it because I suspect I am not the only person who wants something more useful than another opaque score. My hope is that it can help others bring the same clarity to their training decisions—and adapt that guidance to their own bodies, goals, and lives.