MindGuardian is a modular, emotion-aware mental wellness companion built with Flutter. It empowers users to track moods, reflect on meals and workouts, engage in empathetic journaling, and receive personalized support — all powered by Agentic AI.
MindGuardian was developed in two phases:
- Integrated locally hosted Granite 3.1-Dense-2B via
ngrok - Used Flutter + Hive + Firebase Firestore
- Granite provided emotion classification and empathetic responses
- Demo captured showcasing offline + tunneled AI interaction
- Migrated to AWS Lambda (NovaResponder) + API Gateway
- Used S3 for storing mood logs and Firestore for cloud sync
- Integrated Data Prep Kit for training/testing diverse emotional inputs
- Demo captured showcasing real-time cloud-based AI interaction
- ✅ Mood tracking with emotion analysis and journaling
- ✅ Exercise and diet reflection with wellness suggestions
- ✅ StudyBuddy AI for academic support
- ✅ Crisis detection and escalation to support resources
- ✅ Modular tool architecture for scalable wellness features
- ✅ Empathetic responses powered by Agentic AI
MindGuardian supports Theme 2: Mental Health Technology Support:
- Emotion tagging and stress scoring via AI
- RAG-ready architecture for psychological literature grounding
- Modular tools + Data Prep Kit for tailored suggestions
- Agentic AI adapts responses based on user behavior
- Risk score logic triggers escalation to clinician contact
- Empathetic tone and context-aware support
- Flutter, Hive, Firebase Firestore
- Granite 3.1-Dense-2B (local)
- ngrok (secure tunnel)
- Flutter, Hive, Firestore
- AWS Lambda (NovaResponder), API Gateway (with key)
- AWS S3, IAM Roles
- Data Prep Kit
- 🔄 Age-based emotion risk scoring
- 🧠 Voice journaling and speech pattern analysis
- 📊 Clinician dashboard for emotional trends
- 🌍 Multi-age interaction design
To run the app locally:
flutter pub get
flutter run