MindShift: AI Coaching

AI-powered mood tracker and wellness coach that shifts your mindset, one day at a time

Comment

Inspiration

We believe that mental health is one of the most pressing challenges of our time, and we wanted to be part of the solution. With MindShift, our goal is to help people address their mental health in a way that is simple, accessible, and affordable. We envisioned a tool that streamlines the process of self-care and personal growth using empathetic, AI-powered micro-coaching. By focusing on short, effective methods, we aim to make mental health support available to everyone, without the high costs or long, overwhelming sessions. MindShift is our step toward empowering individuals to take control of their mental well-being, one small shift at a time

What it does

MindShift is a personalized AI coaching app that offers: -1–5 minute AI-guided micro-sessions tailored to the user’s mood and goals. -Intelligent journaling with automatic sentiment and theme extraction to help users reflect meaningfully. -Adaptive habit nudges and session sequencing based on user engagement patterns. -Push notifications (via OneSignal) for timely, empathetic nudges. -In-app subscriptions and paywalls (powered by RevenueCat) for premium features. -Realtime sync and authentication (via Supabase) for seamless cross-device continuity.

How we built it

-Frontend: Built with Expo (React Native) for fast Android builds -Backend: Powered by Supabase for authentication, Postgres database, and real-time journaling and user data sync. -AI: Integrated OpenAI Assistant as the base model-powered by GPT-4o and Gemini 2.5 Pro, with prompt orchestration, safety checks, and short context windows for cost efficiency. -Notifications: Leveraged OneSignal for segmented push campaigns and engagement experiments. -Monetization: Implemented RevenueCat SDK for subscriptions, trials, and paywall flows (Shipathon eligibility). -Glue: Lightweight serverless API routes to mediate prompts, enforce rate limits, and ensure safe data persistence.

Challenges we ran into

-Empathy in AI: Tuning prompts to be empathetic while avoiding prescriptive or medical language. -Context Management: Balancing context length to control costs and prevent sensitive information leaks. -Network Resilience: Reconciling real-time client state with Supabase during flaky network conditions. -Risk Management: Designing escalation flows for high-risk signals while minimizing false positives.

Accomplishments that we're proud of

-Delivered a full end-to-end flow: onboarding → micro-session → journaling → paywall (RevenueCat). -Achieved real-time journaling with sentiment and theme extraction, safely persisted in Supabase. -Conducted OneSignal-driven engagement experiments, boosting same-week retention rates. -Enabled rapid iteration with EAS builds and OTA updates, showcasing our progress in #BuildInPublic.

What we learned

-Micro-sessions drive repeat engagement far more effectively than long lessons. -Explicit system prompts and short context windows reduce risky outputs and lower costs. -Privacy controls and data export options significantly enhance user trust. -Push notification timing and message copy matter more than frequency for re-engagement.

What's next for MindShift: AI Coaching

-Expanding AI capabilities to support more nuanced coaching styles and deeper personalization. -Introducing team-based coaching to foster collaborative growth and accountability. -Enhancing privacy features like encrypted journaling and advanced data export options. -Scaling OneSignal experiments to improve engagement across various user segments. -Exploring wearable integrations for real-time mood tracking and habit reinforcement.

MindShift is more than just an app: it’s a movement to make personal growth accessible, empathetic, and effective for everyone.

Let’s shift minds, one micro-session at a time.

Built With

Share this project:

Updates