AlloStatus

Current stress tools simply quantify your stress levels. AlloStatus tells you how much runway you have left before your body stops coping, and the factors that burn through that runway the fastest.

Comment

Our Inspiration

As high school students ourselves who have just recently finished junior year, we've had our fair share of stress and have experienced the negative effects that come with it firsthand. We wanted to create a better way to monitor one's stress that goes beyond simply telling you how stressed you are right now, because the reality of the situation is that by the time you feel burned out, your body has already been in crisis for weeks. We wanted to build something that allows you to take preventative action towards your mental and physical health rather than corrective action.

What It Does

Our platform is inspired by a psychological model known as the General Adaptation Syndrome, which describes how the body sustains a finite resistance buffer under chronic stress. Research shows that heart-rate consistency, sleep consistency, exercise, and perceived stress are all meaningful indicators of how much of that buffer remains. AlloStatus pulls in your wearable data, such as HRV, sleep, and resting heart rate, alongside a short lifestyle questionnaire covering perceived stress, exercise, diet, and social support. It combines these into a single Resilience Buffer Score that reflects how much stress resistance capacity you currently have, and breaks down exactly which factors are helping or hurting. A 30-day trend line shows how your buffer has shifted over time, and an AI-generated insight summarizes your specific pattern in plain English, with ranked recommendations for where to focus first.

How we built it

AlloStatus pulls in wearable data (HRV, sleep consistency, resting heart rate) alongside four lifestyle inputs: exercise, diet, perceived stress, and social support. It combines these into a single 0–100 Resilience Buffer Score modeled on allostatic load indices, weighted against peer-reviewed literature. Rather than telling you how stressed you feel, it shows a ranked breakdown of which factors are draining your buffer most relative to your own 30-day baseline, and pairs each with one concrete action you can take today. A trend line shows how your buffer has shifted over time, and live sliders let you run what-if scenarios in real time to see how changing a habit would move the score.

Challenges we ran into

Our main challenge was turning an abstract scientific idea into a single, reliable number. "Resilience reserve" is grounded in real research, but reducing it to one 0 to 100 score risked becoming either confusing or too much like a medical diagnosis. We chose instead to rank the factors draining a user's reserve rather than predict a burnout date. The statistics were also difficult, since scoring each factor against a user's own baseline meant handling users with little history and measuring sleep timing carefully. We had to work around wearables that are limiting access, since Google Fit is deprecated and Apple Health has no web API. Finally, our goal of a zero-setup experience meant the same engine had to run the same way on the server, in the browser, and in tests, with no database or API keys.

Accomplishments that we're proud of

We are most proud that the score is fully transparent rather than a black box. It uses six factors, each weighted from the research and shown with its citation in the app, so users can see exactly how their reading is built. The engine is pure TypeScript with no database or network, which lets the dashboard sliders recompute the score in real time against the real model. This also gave us a true zero-setup demo: with an empty config and two commands, the full experience runs, including the score, the factor breakdown, the trend, the sliders, and a Gemini coach. We delivered all of this as a deployed product in a single day as a team.

What we learned

Our most important lesson was the value of restraint, since an honest, focused claim proved far more useful than an impressive but unreliable prediction. We also learned a great deal about personal-baseline statistics, including z-scores, fallbacks for new users, and handling time-of-day data. On the engineering side, keeping the scoring model pure made it easy to test and gave us live in-browser scenarios for free, while placing each wearable behind a single interface kept an unstable Google Fit and an API-less Apple Health from affecting the rest of the app. Finally, the Vercel AI Gateway gave us a practical way to deliver a real AI feature without requiring users to bring their own API key.

What's next for AlloStatus

In future iterations, we plan to expand AlloStatus by integrating additional APIs such as Whoop, Terra, and Oura to improve accessibility and data coverage. We also aim to extend the platform beyond individual consumers into educational, workplace, and clinical settings. In so doing, AlloStatus can have a significant impact in multiple different industries. On the research side, we plan to incorporate more advanced biological parameters, including HPA axis sensitivity and autonomic nervous system reactivity. This would allow AlloStatus to create more comprehensive and scientifically grounded assessments of allostatic load and stress resilience.

Share this project:

Updates