MediHub - The Future of Medical Education

Inspiration

Medical students often juggle multiple tools for flashcards, deadlines, research summaries, mnemonics, and case practice. We were inspired to build MediHub to unify these scattered resources into a single AI-powered platform that empowers students to learn smarter—not harder. We wanted to bring the best of modern AI tools into the hands of future doctors in a seamless, secure, and intelligent way.

What it does

MediHub is an all-in-one web platform for medical education. It includes:

• AI-powered summarization of research papers
• Adaptive flashcards with spaced repetition
• A smart deadline tracker for exams, rotations, and applications
• AI mnemonic generator using GPT
• AI patient simulator with voice-based case interactions
• Advanced study analytics and user achievements
• Subscription/paywall system for monetization

Users can upload research papers, generate flashcards and mnemonics, simulate patient cases with voice synthesis, and track their study performance over time—all in one platform.

How we built it

• Frontend: React 18 + TypeScript + Vite + Tailwind CSS + Zustand
• Backend/Database: Supabase (PostgreSQL, Auth, Storage, Edge Functions)
• AI/ML:
  • Transformers.js (DistilBART) for local PDF summarization
  • OpenAI (GPT-3.5) for mnemonic generation
  • ElevenLabs for voice synthesis in the patient simulator
  • Tavus for AI-generated demo videos
• Payments & Subscriptions: RevenueCat + Stripe integration
• Development Tools: bolt.new for rapid prototyping, Framer Motion for animations

We used a modular architecture, feature gating via custom components, and mock/test modes to simulate Pro features locally. Data privacy and usability were key priorities throughout.

Challenges we ran into

• Summarization accuracy: Balancing speed, privacy, and performance was tough when deploying Transformers.js locally.
• Voice synthesis integration: Making ElevenLabs work reliably within browser constraints required careful handling of audio playback and API throttling.
• Subscription logic: Syncing RevenueCat with Supabase in real-time using Edge Functions and webhooks took careful planning.
• Scalability: Designing for both mock/dev mode and production-grade paywall required dynamic feature toggling and secure data handling.

Accomplishments that we're proud of

• A fully functional, multi-featured platform integrating AI, user accounts, payments, and advanced analytics in under one cohesive UX.
• Real-time research summarization using local models—no server inference needed!
• Seamless voice-driven patient simulator that mimics real medical scenarios.
• A robust subscription/paywall system with mock mode, webhook syncing, and feature gating.
• Built entirely using modern web technologies—React + Supabase + Vite + Zustand—and launched from scratch.

What we learned

• How to efficiently integrate multiple AI services (OpenAI, ElevenLabs, Tavus) within a privacy-first, performance-conscious application.
• Supabase is incredibly powerful for rapid prototyping and production if used with well-defined access policies.
• RevenueCat is not just for mobile—its web support opens the door to freemium monetization for SaaS tools.
• UX design is as critical as backend logic in educational tools—students need clarity, responsiveness, and joy in the interface.

What's next for Medihub - The future of medical education

• Mobile App Launch: Expand the platform to iOS and Android using React Native or Swift/Flutter.
• AI Study Coach: Personalized recommendations, performance insights, and motivational nudges.
• Community Sharing: Flashcard decks, mnemonics, and patient cases can be shared, rated, and improved by the community.
• Live OSCE Simulations: Real-time, timed clinical roleplay scenarios with AI examiners.
• Institutional Version: Offer a dashboard for medical schools to track cohort performance and generate training insights.

Built using:
React, TypeScript, Vite, Tailwind CSS, Supabase, Framer Motion, OpenAI, Transformers.js, ElevenLabs, Tavus.ai, RevenueCat, Stripe, bolt.new

Made with ❤️ by the MediHub team.

Built With

• elevenlabs-api
• eslint
• framer-motion
• lucide-icons
• openai-api
• postgresql
• prettier
• react
• recharts
• revenuecat
• stripe
• supabase
• supabase-edge-functions
• tailwind-css
• tavus-api
• transformers.js
• typescript
• vite
• zustand