Predict Health Toilet (PHT) is a health platform that manages users and measurement data from smart toilet sensors, and delivers dehydration analysis through backend services, a web application, and machine learning models.
This project is research-only open source and released “not for clinical use”. It can ONLY be used for:
- Academia
- Prototyping
- Benchmarking
Subprojects
apps/monorepo: REST API (FastAPI), web application (React/Vite), database migrations, and Docker-based infrastructure for local and production deployment. See apps/README.md.ml/workspace: machine learning pipelines for dehydration risk, including population-level modeling and urine biomarker inference, with Azure ML pipeline definitions and utilities. See ml/README.md.
Quick Links
- apps/README.md (platform apps, services, and infrastructure)
- apps/docs/api/README.md (API endpoints, RBAC, auth)
- apps/docs/database/README.md (schema, relationships, indexes)
- apps/docs/infrastructure/README.md (local + production infra)
- apps/docs/webapp/README.md (web app notes)
- ml/README.md (machine learning workspace and pipelines)