MedGPT is a state-of-the-art AI-powered medical diagnostic tool designed to provide rapid, accurate second opinions for critical conditions like stroke and breast cancer. By leveraging advanced neural networks and large language models (Falcon AI), MedGPT ensures reliable validation of patient conditions through a hybrid analysis of neuro-imaging (MRI) and clinical questionnaires.
- AI-Powered MRI Analysis: Real-time diagnostic assistance using deep learning models to identify stroke indicators with 99.2% precision.
- Dynamic AI Pipeline: A seamless workflow from raw data input to clinical insights, visualized in a high-tech dashboard.
- Empathetic AI Chatbots: 24/7 mental health, caregiver support, and rehabilitation guidance powered by Falcon AI.
- Cross-Platform Readiness: Responsive, glassmorphism-based UI designed for smartphones, tablets, and medical workstations.
- Medication Reminders: Integrated scheduling system to ensure adherence to treatment protocols.
| Component | Technology |
|---|---|
| Frontend | Next.js 14 (App Router), TypeScript, Tailwind CSS |
| Animations | Framer Motion |
| Backend/Auth | Supabase (PostgreSQL, SSR Auth) |
| AI/ML | TensorFlow (Image Recognition), Falcon AI (NLP) |
| UI Components | Radix UI, Lucide Icons, Sonner |
MedGPT follows a modern AI pipeline:
- Data Ingestion: MRI scans (DICOM/PNG) and clinical data are securely uploaded.
- Preprocessing: Image normalization and NLP parsing of symptoms.
- Inference Layer:
- Vision Model: TensorFlow-based CNN identifies neurological anomalies.
- Language Model: Falcon AI processes clinical history and provides support.
- Reporting: Interactive glass-morphic dashboard delivers risk assessment and next-step recommendations.
- Node.js 18+
- Supabase Account
- Clone the repository:
git clone https://github.com/itxtalal/medgpt.git
- Install dependencies:
pnpm install
- Set up environment variables:
Copy
.env.exampleto.env.localand add your Supabase credentials. - Run the development server:
npm run dev
- Muhammad Talal Jamil - GitHub
- Ehtisham Khan - GitHub
- Areesa Ashfaq
- Sandra Ashipala - GitHub
- Caner Akturk
This project is licensed under the MIT License - see the LICENSE file for details.
Built with ❤️ for the AI Hackathon 2026.