ARNOVA HEALTH.AI

Inspiration

Indonesia is approaching a once-in-a-lifetime demographic dividend to achieve the Golden Indonesia 2045 vision. However, millions of young people the spearhead of our nation's productivity face severe mental health challenges like stress, anxiety, and burnout that often go unnoticed until they reach a crisis point. We were inspired to create a proactive, stigma-free early intervention platform to prevent these emotional struggles from impacting productivity and economic burden. We blended modern clinical psychology with universal spiritual values to make the healing approach deeply human-centered and inclusive.

What it does

ARNOVA HEALTH.AI is not just a standard chatbot; it is an AI Observation Engine. It acts as a trusted Reflection Companion that guides users through their emotional journey. Our core features include daily mood assessments, AI-powered intelligent journaling, and personalized recovery recommendations. The AI observes user metrics over time (such as mental energy levels and burnout risk) to provide increasingly accurate and targeted supportive guidance without replacing professional psychologists.

How we built it

We built ARNOVA HEALTH.AI using a modern full-stack architecture designed for rapid prototyping during the hackathon.

• Frontend & Backend: Fully built with Next.js (App Router) to keep the interface development and API Routes in one clean repository.
• AI Reasoning Engine: We utilized the Gemini 2.5 Flash API for its blazing-fast text processing and reliable JSON object generation.
• Database & Authentication: Supabase powers PostgreSQL, authentication (Google OAuth and email/password), Row Level Security, and file storage.
• UI/UX: The interface was designed mobile-first using Tailwind CSS, featuring glassmorphism effects and a calming pastel color palette.

Accomplishments that we're proud of

• Successfully building an AI Observation Engine with structured memory, ensuring the AI doesn't get "amnesia" and can remember the user's context from previous days.
• Designed a unified Personal Recovery Planner, allowing AI recommendations to stay consistent across Dashboard, Home, and Insights.
• Implemented fallback mechanisms so the application remains functional even when AI generation is unavailable.
• Designing a highly responsive, beautiful UI/UX that feels like a premium wellness app, complete with a complex Curved Bottom Navigation Bar.

What we learned

• We learned that designing a helpful AI experience is more challenging than simply integrating an LLM. Much of our time was spent refining prompts, deciding when the AI should regenerate recommendations, and ensuring its responses remained consistent across the application.
• Sometimes, simplifying a complex relational database into a flat table is the smartest decision to ensure a smooth live demo on the competition stage.

What's next for ARNOVA HEALTH.AI

Building a truly empathetic, human-centered AI requires time, care, and rigorous testing—things that are impossible to perfect within a short hackathon weekend. Because our current AI training and knowledge base injection haven't reached their maximum potential yet, we strongly believe this project is highly worth developing beyond this competition. The problem we are solving is deeply crucial for our generation's future. Moving forward, we will focus on:

• Deepening the AI's Knowledge: Enriching the psychological and spiritual observation datasets in our vector database so ARNOVA's responses become even more comprehensive, safe, and accurate. Retrieval-Augmented Generation (RAG): Build a structured knowledge base so the Recovery Planner can retrieve trusted psychological and wellness resources before generating recommendations.
• Wearable Integration: Connecting with smartwatches and push notifications to monitor users' biological data (like heart rate variability) in real-time.
• Gamification: Building an achievement system so users feel more encouraged to consistently maintain their mental health routines.

Built With

• ai
• api
• next.js
• supabase
• tailwind