A wearable-driven platform that predicts diabetes complications before they become irreversible.
Diabetes management is often reactive. Patients lack visibility into the silent progression of dangerous complications — retinopathy, nephropathy, neuropathy, and diabetic foot — while healthcare providers suffer from a lack of continuous, real-time data to intervene early. dabetai solves this by providing a preventive ecosystem that fuses real-time biological data from wearables with clinical oversight and AI-powered risk prediction, shifting diabetes care from reactive to proactive.
┌─────────────────────────────────────────────────┐
│ Mobile App (Patient Hub) │
│ React Native + Expo + Tailwind CSS │
└───────────────────────┬─────────────────────────┘
│ HTTPS / JWT
┌───────────────────────▼─────────────────────────┐
│ Core API (NestJS) │
│ PostgreSQL + Prisma + JWT Auth │
└────┬──────────────────────┬─────────────────────┘
│ │
▼ ▼
┌──────────────┐ ┌──────────────────┐
│ Web Portal │ │ AI Inference │
│ (Medical) │ │ API │
│ Angular 19 │ │ FastAPI + │
│ Tailwind │ │ scikit-learn │
└──────────────┘ │ XGBoost │
└────────┬─────────┘
│
┌────────▼─────────┐
│ AI Models Core │
│ Python/PyTorch │
└──────────────────┘
| Repository | Purpose | Stack | Status |
|---|---|---|---|
| mobile-app | Patient hub for glucose monitoring, wearable sync, and AI risk alerts | React Native 0.79, Expo 53, Tailwind CSS | ✅ Active |
| web-app | Medical portal for remote patient oversight and clinical insights | Angular 19, Tailwind CSS | ✅ Active |
| api | Core REST API: auth, users, medical data, AI communication | NestJS 11, PostgreSQL, Prisma | ✅ Active |
| ai-api | AI inference API for real-time complication risk prediction | FastAPI, Python 3.11, MongoDB | ✅ Active |
| ai-models | ML pipelines: training, evaluation, and serialization of predictive models | Python, scikit-learn, XGBoost, PyTorch | ✅ Active |
| landing | Educational landing page presenting the ecosystem | Astro, Tailwind CSS | ✅ Active |
- Patients use the Mobile App to log glucose, food, medication, and physical activity — or sync with wearables (CGMs) for real-time biomarkers like heart rate and sleep quality
- Healthcare providers connect via the Web Portal to monitor patients remotely, review clinical trends, and receive AI-driven alerts
- The AI Inference API processes patient data through trained machine learning models to forecast the risk of specific complications: retinopathy, nephropathy, neuropathy, and diabetic foot
- Both patients and doctors receive early warnings, enabling timely intervention before complications become irreversible
Published paper: "Prevención de Riesgos de la Diabetes Mediante una Plataforma Inteligente de Monitorización y Predicción de Complicaciones con Inteligencia Artificial" — Read the full paper
- End-to-end JWT authentication
- Role-Based Access Control (RBAC)
- Encrypted medical data storage
- Secure wearable device integration
We follow a strict PR-based workflow. Check each repository's CONTRIBUTING.md for guidelines.
Built with ❤️ by the dabetai team · 2026
