AI Products
Hoxa
A fitness product for people who want structure and progress without the usual gamified noise.
2026 / AI Products / SwiftUI / SwiftData / FastAPI / OpenAI / TypeScript
Quick read
A clear breakdown of the product, the reasoning behind it, and the implementation choices that matter.
- Year
- 2026
- Category
- AI Products
- Status
- In design
Overview
Hoxa is a fitness product for people who want to get stronger and move more confidently without the streaks, badges, and guilt-tripping most fitness apps lean on. It brings running, strength, mobility, balance, and recovery into one training loop, with a FastAPI backend and an adaptive engine that responds to adherence, effort, and context while explaining its decisions in plain language.
Why it exists
A lot of fitness products lose people before training becomes useful because they assume the user already feels like an athlete. The language, design, and intensity cues often create pressure before they create confidence.
How it works
Hoxa uses a plainer tone and an adaptive engine that responds to real signals — fatigue, missed sessions, a bad week — without treating any of it as failure. When the plan changes, it says why.
Built with
SwiftUI / SwiftData / FastAPI / OpenAI / TypeScript
What I focused on
- Designing a multi-mode training system covering running, strength, mobility, balance, and recovery.
- Building an adaptive training engine that adjusts plans based on adherence, effort, and context signals.
- Keeping the product tone measured and non-performative to lower friction for beginner and intermediate users.
- Planning integrations with Apple Health, Apple Watch, and Garmin for richer context over time.