Mindful Buddy

A virtual AR pet that guides college students to better mental health with AI‑driven journaling, meditation, and supportive interactions.

Comment

Mindful Buddy

Inspiration

1 in 3 college students lives with a diagnosable mental health condition. Research shows that interacting with animals triggers the release of oxytocin and dopamine, reducing cortisol levels and anxiety. However, most dorms don’t allow pets. In recognition of National Pet Day, we sought to bridge that gap. A virtual pet could deliver some of that same comfort, while actively guiding students toward better mental‑health habits.

What we learned

Technical

  • Flutter & Dart – cross‑platform mobile development.
  • ARKit (via a Flutter AR plugin) – augmented‑reality pet experience.
  • Supabase – real‑time database, authentication, and social features.
  • ElevenLabs – voice generation for meditation guidance.
  • Gemini API – conversational AI for journaling prompts.

These tools came together to create a responsive journaling and meditation experience.

Design

  • Empathetic dialogue: The pet validates feelings, offers gentle encouragement, and suggests coping strategies in a non‑judgmental tone.
  • Positive reinforcement: Completing a journal entry or meditation session triggers calm animations and soothing sounds, reinforcing self‑care habits.
  • Safe space: Conversations are designed with clear privacy cues and easy opt‑outs, helping users feel comfortable sharing.
  • Thoughtful UI: Colors, animations, and haptic feedback are selected to promote calm and reduce stress.
  • Community kindness: Friends can send supportive notes and virtual treats, fostering a network of encouragement.

How we built it

Layer Technology
Frontend Flutter (Dart) with a Flutter AR plugin (ARKit) + haptic feedback
Backend Supabase – auth, user data, real‑time social features
AI Gemini API – interactive journaling prompts
ElevenLabs – audio‑based meditation guidance

Wellness‑tracking formula

The wellness tracker mirrors a GitHub contribution graph. Let

  • J = number of journal entries in a given week
  • M = meditation sessions completed in that week

The weekly wellness score W is calculated as

$$ W = \alpha J + \beta M $$

where α and β are weighting coefficients tuned to reflect the relative impact of each activity on mood outcomes.

Challenges

  • AR performance: Making the pet feel alive without overloading the device required heavy optimisation of render pipelines and haptic timing.
  • Scheduling logic: Crafting a reliable system for weekly tracking, reminders, and streak calculations proved tricky and went through several iterations.

Mindful Buddy demonstrates how a blend of AR, AI, and thoughtful design can bring comfort and healthy habits to students—especially when real pets aren’t an option.

Built With

Share this project:

Updates