This project introduces a system that utilizes the Flower framework, along with the ESP32 microcontroller and the TinyDB database, for stress classification. The system collects and processes real-time biomarker data, enabling local model training on edge devices.