ProtoFitSenseAI

ProtoFitSenseAI exists because no fitness app could handle my reality. After a decade of powerlifting, I had a collection of injuries, imbalances, and bad habits — hip issues, terrible posture from building all day at a computer, movement patterns that needed unlearning. My training data was scattered across five different apps. My injury history lived in hip X-rays, MRIs, and notes from six different physios. None of it talked to each other.

I kept going to ChatGPT, explaining my full context from scratch every time — injuries, goals, equipment available, what I did last week. And I thought: this should be one app that knows me and evolves with me.

Why It Has to Be Adaptive

My needs change constantly. My hip is improving — I need harder progressions. My posture is getting better — time to shift focus. I’m jet-lagged after a flight — adjust the session. I skipped three days — program a proper reentry instead of throwing me back at full volume. I want to start running again — factor that into my lifting recovery. I switched gyms and have different equipment — adapt.

No static 12-week program can handle this. A good human coach could, but most people don’t have one. ProtoFitSenseAI is my attempt to build that coach — one that holds my full context (medical history, training logs, goals, life circumstances) and makes intelligent session-by-session decisions.

Technical Approach

The system takes structured inputs — exercise history, load progression, recovery signals, schedule constraints — and feeds them to an LLM with fitness programming principles baked into the prompt architecture. The output is a concrete training plan: exercises, sets, reps, and load recommendations that adjust based on how your last sessions went.

For example, if you hit all your squat sets last week and rated recovery as high, it’ll progress the load. If you slept poorly and flagged fatigue, it’ll auto-regulate volume down. It’s practical AI applied to a deeply personal, context-dependent problem.

Current Status

I use this every single day. It’s my own training partner — functional, evolving, and getting smarter as I feed it more data and refine the programming logic. It’s not a polished consumer app yet. It’s a working tool that I’m building in public, iterating based on real usage with a user base of one very demanding user: me.

The long-term goal is to help anyone with complex training needs — people with injuries, imbalances, or goals that don’t fit neatly into a cookie-cutter program — train smarter without needing a $200/month coach.