FeelBit
Mood tracking and reflection, built as a JavaFX desktop app and rewritten for the web.
Product
FeelBit lets a user record a daily mood, associate it with common triggers, and add a private journal entry. Mood analytics surface patterns over time, while cached Gemini-powered wellness suggestions provide optional support for reflection.
What I built
FeelBit was developed with a four-person Scrum team as a desktop application using JavaFX, MongoDB, and the Gemini API. I then rebuilt it for the web as a React client and Spring Boot API backed by PostgreSQL and Flyway migrations, with a Docker Compose setup that runs the system consistently across environments.
Engineering focus
- Model mood, trigger, journal, and analytics data behind a Spring Boot service boundary
- Use one-hour JWT access tokens and rotating seven-day refresh tokens stored as SHA-256 hashes
- Protect credentials with BCrypt and automatically restore authenticated sessions
- Cache AI-powered wellness suggestions to reduce repeated model work
Stack
Java, Spring Boot, React, PostgreSQL, Flyway, Docker Compose, JWT, BCrypt, and Google Gemini API. The original desktop version uses JavaFX and MongoDB.