EverWell

EverWell

Inspiration

We were inspired by conversations with our grandparents and other elderly individuals who often spoke about loneliness, lack of purpose, and cognitive decline as some of the hardest challenges of aging.
We wanted to build a compassionate, easy-to-use solution that helps them not only track their cognitive and emotional health, but also reconnect with loved ones and peers in a meaningful way.

Our Product

Our app provides an interactive dashboard where elderly users can track their progress across both mental health and cognitive function using short, engaging games and reflection activities.

Key Features

Daily Brain & Mood Games: Exercises inspired by the Stroop Test and List Recall games to assess attention, memory, and reaction speed.
Progress Tracking Dashboard: Visual summaries that show improvement trends over time.
Connection Mode: A “social circle” feature that lets users share results and messages with family or friends.
Multiplayer Games with LLM Facilitation: Games can be played cooperatively or competitively, with a language model acting as a mediator or host, making the experience conversational and friendly.
Accessibility for the Visually Impaired: Integrated Fish Audio APIs convert tasks and feedback into speech so users with limited vision can participate equally.

How We Built It

We developed the app using the MERN stack:

Frontend: React Native for cross-platform mobile accessibility (iOS and Android).
Backend: Node.js and Express for API handling and real-time communication.
Database: MongoDB for secure, scalable storage of user data and performance metrics.
Audio Layer: Fish Audio API for speech-based interactions and accessibility.
Chat + Multiplayer: JanitorAI API for enabling conversational game facilitation and social connection.

We also incorporated a token-based authentication system (JWT) for secure user sessions and integrated LLM-driven prompts that adapt the difficulty of tasks based on performance trends.

Challenges We Faced

Accessibility Design: Ensuring games were equally usable by both sighted and visually impaired users required extensive experimentation with audio feedback and timing.
LLM Integration: Calibrating the language model to act as a friendly “game host” without over-responding or confusing users was a significant UX challenge.
Data Visualization: Representing cognitive data trends in a way that is intuitive and not overwhelming for older users took several design iterations.

What We Learned

We learned how technology can bridge generational gaps when designed with empathy and accessibility in mind. We also deepened our understanding of integrating audio-based AI models, handling real-time multiplayer state management, and designing for users who are often overlooked by mainstream tech.

Future Vision

We plan to expand the app with:

Personalized difficulty calibration using reinforcement learning and adaptive learning features.
Integration with wearables (e.g., Apple Watch, Fitbit) to combine physical and cognitive metrics.
AI-driven health insights for early detection of cognitive decline.