Inspiration
Middle-aged women in India often prioritize family over their own health, leading to stress, poor lifestyle habits, and emotional neglect. We wanted to build a solution that not only tracks wellness but also involves family support in a meaningful and culturally relevant way.
What it does
WellNest is a maternal wellness app with two connected experiences: Mother and Family.
For mothers, it provides:
Conversational daily check-ins Mood and habit tracking Emotion-aware feedback from Nesti (the in-app companion) Wellness reports with trend charts and downloadable PDF summaries Sanctuary and journaling spaces for reflection For families, it provides:
Family dashboard visibility based on privacy settings Care Pulse messaging to send gentle encouragement Love Archive to preserve supportive messages Risk-aware alerts when support may be needed AI capabilities include:
BERT-based emotion detection from free-text notes LDA topic discovery from journal entries Random Forest wellness risk prediction with explainable factors and recommendations
How we built it
We built WellNest as a multi-service architecture:
Frontend: React + TypeScript + Vite + Tailwind + component-driven UI Backend API: Spring Boot with authentication, data persistence, and real-time communication support ML layer: FastAPI microservice exposing BERT, LDA, and risk prediction endpoints Data flow: secure API communication, role-based views, and AI-assisted insight rendering in dashboards The product experience was designed around emotional safety, clarity, and actionable support rather than clinical complexity.
Challenges we ran into
Integrating three different stacks (React, Spring Boot, Python ML) while keeping API contracts consistent Making AI insights useful without being overwhelming or alarmist Handling privacy controls so mothers can choose exactly what is shared Balancing real-time family support with autonomy and emotional sensitivity Translating technical outputs (scores, labels, confidence) into empathetic language
Accomplishments that we're proud of
End-to-end working platform with role-based user journeys Explainable AI insights integrated directly into daily wellness flows Human-centered features like Care Pulse, Love Archive, and privacy levels Reporting experience with trend visualizations and exportable PDF summaries A cohesive product that combines emotional design with production-style architecture
What we learned
Mental wellness products need trust-first UX, not just prediction accuracy Explainability and phrasing are as important as model outputs Cross-stack integration succeeds when API schemas are clean and explicit Privacy controls should be first-class features, not post-launch add-ons Family support tools work best when they are proactive, lightweight, and kind
What's next for WellNest
Personalized interventions based on longitudinal behavior patterns Multi-language emotional intelligence support Smarter alert threshold tuning to reduce false positives Clinician and counselor integration pathways Voice-first check-ins and improved accessibility features Deployment hardening, monitoring, and larger-scale user testing
Built With
- axios
- fastapi
- framer-motion
- h2
- java-17
- jwt
- mysql
- numpy
- pandas
- python
- pytorch
- react
- recharts
- scikit-learn
- shadcn/ui
- spring-boot
- spring-data-jpa
- spring-security
- tailwind-css
- transformers
- typescript
- uvicorn
- vite
- websocket
Log in or sign up for Devpost to join the conversation.