MindMosaic

Voice stress analysis
Real time monitoring on basis of voice input
Text sentiment analysis, daily challenges to make you feel better.

MindMosaic is an AI-powered solution designed to detect early signs of hidden mental health decline in corporate employees and students. It addresses the growing issue of silent burnout, anxiety, and depression, which often go unnoticed until a crisis occurs.

The system uses AI to analyze both voice and text inputs to capture subtle emotional shifts. During video meetings or journaling, MindMosaic passively tracks sentiment patterns to understand the user's mental state over time. Based on these patterns, it provides timely nudges and wellness prompts through augmented reality (AR) avatars, encouraging reflection or breaks when needed.

To maintain privacy, MindMosaic uses federated learning. This ensures that user data never leaves their device, while the overall model continues to improve through local training. The platform also includes a real-time dashboard that shows personalized mood trends and mental wellness analytics.

For organizations and educational institutions, MindMosaic offers an aggregated view of group wellness data—without revealing individual details. This helps HR teams and counselors make informed decisions and intervene early when negative patterns are detected at scale.

Technically, the solution integrates natural language processing, emotion detection APIs, and privacy-first design. It is built using a modern web stack and can integrate with productivity tools like video conferencing platforms and calendar apps.

MindMosaic has the potential to prevent mental health crises by catching early signals and offering actionable support. With its non-intrusive design and strong privacy measures, it creates a safe space for proactive mental wellness.

Built With

ai-powered-sentiment-apis-(nlp-&-speech-to-text)
and
federated
framer-motion
learning
lite
lucide-icons
or
react
recharts
tailwindcss
tensorflow
with

Submitted to

HackForge

Created by

I led the end-to-end development of MindMosaic, an AI-powered web application designed to detect early signs of mental health decline in corporate employees and students. My key responsibilities included:

Ideation & Problem Framing: Conceptualized the project around workplace and academic stress, focusing on early mental health intervention using AI.

Frontend Development:
Built a clean, responsive, and human-centered user interface using React, TailwindCSS, and Framer Motion. Designed multiple interactive sections including the Dashboard, Journaling Interface, and Alerts Panel.

Data Visualization:
Integrated Recharts to create dynamic visualizations for tracking weekly mood trends and sentiment over time.

Sentiment Analysis Integration:
Implemented hooks for connecting text and (future) voice-based sentiment analysis, enabling real-time mental state monitoring.

Gamified Wellness System:
Designed the logic and UI for wellness nudges using AR-inspired prompts to promote breaks and self-reflection.

Privacy-First Architecture:
Proposed the use of Federated Learning using TensorFlow Lite or PySyft to ensure user data never leaves the device.

Deployment:
Deployed the application live using Netlify, making it accessible at https://gorgeous-dragon-429d3c.netlify.app.

Presentation & Branding:
Crafted the project tagline, refined the UI tone to maintain a professional and non-AI-gimmicky appearance, and ensured all components aligned with a real-world mental wellness product.

Honey Priya

Updates

Honey Priya started this project — Jul 08, 2025 03:36 AM EDT

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.