CURA.AI

Practice medicine with realistic AI patients.

Comment

Inspiration behind CURA.AI

As medical students, we lived the "theory-practice gap" every day. We spent years memorizing symptoms and protocols, yet the first time we faced a patient, the fear of making a clinical error was overwhelming. We realized that the current medical education system lacks a "safe sandbox"—a place where students can fail, learn, and refine their clinical reasoning without risk to human life.

Our inspiration is based on three core pillars:

  1. The "Safe to Fail" Zone: Transforming the paralyzing fear of medical mistakes into a risk-free, creative learning playground.
  2. Medicine Without Borders: Ensuring high-fidelity simulation isn't a luxury for elite hospitals, but a tool available in every student’s pocket.
  3. The Clinical Simulator (Dynamic Personas): Moving beyond static Q&A to create living simulations where every AI patient has a unique personality, specific age, and dynamic health states (stable vs. critical) that evolve based on your decisions.

Our Global Impact (SDGs)

CURA.AI stands at the vital intersection of SDG 3 (Good Health and Well-being) and SDG 4 (Quality Education). While our platform leverages cutting-edge educational technology to transform medical training (Target 4.4), our ultimate mission is to fulfill Target 3.c by significantly strengthening the global health workforce. By providing a risk-free simulation environment, we are not just teaching students; we are reducing future clinical errors and ensuring safer, higher-quality healthcare outcomes for communities worldwide.

What it does

CURA.AI is an immersive clinical simulator that transforms medical education into a dynamic, high-stakes experience where every decision counts.

Tailored Learning Paths

  • Specialist vs. Student Modes: The platform adapts to the user. Specialists receive high-complexity cases, while students access a broad range of specialties (Cardiology, Pediatrics, Orthopedics, etc.) to build a well-rounded foundation.

Interactive Diagnosis & Reasoning

  • Beyond Multiple Choice: Students must justify their clinical reasoning, ensuring a deep understanding of the "Why" behind every decision.
  • Smart Clinical Guidance: Integrated Intelligent Hints and Recommended Tests guide users through complex investigations like ECGs or specific lab panels.

Comprehensive Insight Reports

  • Deep Performance Analysis: Every case concludes with a detailed report featuring an Accuracy Score and a breakdown of "Correct vs. Incorrect" actions.
  • Portable Documentation: Summaries are available as Downloadable/Printable PDF Reports, perfect for clinical portfolios.

Emergency Mode: The Ultimate Challenge

  • High-Pressure Simulation: A dedicated mode with a 10-minute countdown timer to test speed and accuracy under stress.
  • The "Unknown Specialty" Factor: The specialty is hidden, forcing users to perform rapid differential diagnoses based purely on clinical presentation and raw lab results.

How we built it

CURA.AI is built with a high-performance full-stack architecture, focusing on stability and medical accuracy.

  • Frontend & UX Design: Built with React.js (via Vite) for ultra-fast performance. We translated Figma designs into custom Vanilla CSS, featuring a glassmorphic interface with professional purple and baby blue gradients.
  • AI Engine & Backend: Powered by Node.js and Express.js. Our modular engine supports multiple LLMs like Gemini and Ollama. We used advanced System Prompt Engineering and integrated a library of 10 pre-built cases to solve latency issues during demos.
  • Data & Optimization: We utilized Supabase (PostgreSQL) for secure auth and session management. We built a Guest Mode using secure UUIDs, optimized with React.useMemo to prevent infinite rendering loops.
  • Deployment: The entire ecosystem is unified and deployed via Vercel using a serverless architecture for scalable global access.

Challenges we ran into

  • Combating AI Hallucinations: Implementing rigorous prompt engineering and knowledge-base constraints to ensure the AI stays 100% medically accurate .
  • Bridging the Med-Tech Gap: Translating complex "doctor-speak" into clean code and structured data models.
  • The Design Challenge (UX/UI): * Visual Harmony: Using a soothing palette of Light Purple and Baby Blue to reduce anxiety.
    • Information Hierarchy: Organizing dense medical data (labs, vitals, chats) into a clean, "frictionless" layout.

Accomplishments that we're proud of

  • Seamless Med-Tech Integration: Merging clinical logic with AI to create patients that feel "alive" and consistent.
  • Adrenaline-Pumping Emergency Mode: Successfully building a high-stakes timer system that simulates real ER pressure.
  • End-to-End User Journey: Developing a complete ecosystem—from patient greeting to a downloadable PDF report.
  • An Engaging UI design: Creating a unique vibe experienced by integrating animated breathing robot video and Calming color pallete

What we learned

  • Med-Tech Synergy: The incredible potential when medical expertise and software engineering speak the same language.
  • The Precision of AI Prompting: Taming AI to be a realistic patient requires deep medical knowledge and specific logic.
  • The Beauty of Iteration: Realizing that a great project is built through constant testing and refining.

What's next for CURA.AI

The current website is a demo version, and our roadmap includes:

CURA.AI Future Roadmap

  • Advanced AI Refinement: Fine-tuning models for even more complex, multi-systemic conditions.
  • Mobile Expansion: Dedicated Android and iOS apps for "a clinical ward in your pocket".
  • The Rare Disease Project: Integrating specialized databases for symptoms often missed in traditional training.
  • Visual Diagnostic Integration: Incorporating medical imaging (X-rays, CT scans, MRIs) directly into the AI diagnostic flow.
  • Global Accessibility: Adding multi-language support (Arabic, French, Spanish, etc.) to reach every medical student worldwide.

Built With

Share this project:

Updates