Inspiration

As college students, we’ve seen how academic pressure, personal expectations, and everyday stress can pile up. Finding the time and space to understand what we’re feeling can be difficult. We wanted to build something that makes reflection and everyday self-care easier to approach.

That idea became Unwind, a privacy-first mental wellness platform that brings journaling, wellness tracking, AI-assisted conversations, and community support into one place.

What it does

Unwind gives users tools to reflect on their day, understand their habits, and connect with others.

  • Private journaling with PIN protection to add another layer of access control.
  • Wellness trackers for mood, sleep, habits, energy, and water intake.
  • An AI companion for supportive conversations and reflection.
  • Community features including posts, comments, and real-time chat.
  • DASS-21 self-assessments with downloadable reports.
  • Moderation and reporting tools to help manage community activity.

The platform focuses on everyday mental wellness and self-awareness. Its tools do not provide clinical diagnoses.

How we built it

We built the frontend with React and Vite, the backend with Node.js and Express, and used PostgreSQL on Neon to store application data. Socket.IO powers real-time community messaging.

For authentication, we implemented email verification, Google OAuth, and JWT-based sessions with refresh tokens stored in HTTP-only cookies. We also added input validation, rate limiting, role-based access controls, and administrative audit logs.

The frontend is deployed on Vercel, while the backend runs on Render. We use GitHub Actions for CI/CD, UptimeRobot for availability monitoring, and PostHog to understand how people use the platform.

Challenges we faced

One of our biggest challenges was managing sessions across protected pages and real-time connections. Expired tokens, socket disconnections, and navigation issues showed us how closely authentication and frontend state need to work together.

We also faced database usage limits and hosting constraints. These pushed us to investigate resource consumption, improve monitoring, and think more carefully about reliability.

Privacy was another important design consideration. Journals and wellness records are personal, so we had to consider who could access each feature and how to keep private activities separate from community interactions.

What we learned

Building Unwind taught us how much work continues after a feature runs locally. Deployment, authentication, database management, monitoring, and user feedback all became part of our development process.

Testing with 36+ participants helped us identify confusing interactions and issues we had missed during development. We learned to look beyond whether a feature works and consider whether someone can understand and use it comfortably.

We also learned to be careful with the language around mental wellness and AI. Every feature needs a clear purpose, understandable limitations, and claims that match what we have actually tested.

What’s next

We’re working toward more contextual AI-assisted reflection, improved performance, and a structured campus pilot. We also want to collaborate with psychology professionals to review the experience and guide future development.

Our goal is to keep improving Unwind into a practical space where people can pause, reflect, and build a more consistent self-care routine.

Built With

Share this project:

Updates

Submission history