Check my Mental

• 

  Main dashboard

• 

  Luna AI

• 

  Check in emotion tags

• 

  Check in scale

Project story

The idea came from a really personal place. During my first year at Mac, I went through a rough time where I felt completely isolated. My family was hours away, my friends were dealing with their own stress, and I didn't want to burden anyone with my problems. I'd stay up late just spiraling, wishing someone would check in on me. I realized I wasn't alone as so many students related as mental health resources are either limited, have long wait times, or feel intimidating to access. I wanted to build something that could be that first step to supporting my peers.

Building this taught me way more than I expected. I had to carefully design conversations that feel genuine without overstepping into territory that should be handled by real professionals. On the technical side, implementing real time chat while maintaining conversation context was definitely a challenge. Managing conversation context across multiple check ins was tricky since I needed the AI to remember past discussions without the token count going super over.

I built the frontend with React and TypeScript, focusing on a clean, calming interface. The backend runs on Node.js with Express, handling user authentication, scheduled check-ins through node cron, and all the Gemini API calls. PostgreSQL stores user profiles, mood history, and conversation logs so the AI can reference past conversations and spot trends over time. Looking back, I spent a lot time refining crisis detection keywords and making sure emergency resources were always visible.

In the end, a cool prototype was developed and I even had to test it with my friends.

Built With

• express.js
• gemini
• jwt
• node.js
• react
• typescript