Clyra - AI Patient Simulations for Medical Students

Medical students get little real patient time early on. Our AI powered virtual patient platform gives them unlimited, on-demand practice diagnosing, treating, & communicating with realistic patients

Comment

Inspiration

Medical students spend less than 15 minutes per week interacting with real patients in their first two years of training — yet mastering bedside manner and diagnostic reasoning requires hundreds of hours of real-world experience. We wanted to bridge that gap by creating Clyra, an AI-powered virtual patient platform that provides realistic, on-demand clinical encounters.

What It Does

Clyra simulates natural, lifelike patient interactions using generative AI. Students can:

  • Diagnose and treat virtual patients across diverse cases.
  • Practice communication and empathy through realistic dialogue.
  • Receive instant feedback on diagnostic accuracy, empathy, and communication style.

The goal: make clinical training accessible, scalable, and personalized.

How We Built It

We combined multiple technologies to create a responsive, emotionally intelligent AI patient experience:

  • LLMs (Claude API) for realistic dialogue generation.
  • Voice interfaces using text-to-speech and speech recognition.
  • Scenario orchestration through a case management backend built with Python + FastAPI.
  • Frontend using HTML, connected to a real-time chat and voice engine.
  • Data storage for session tracking and analytics.

Challenges We Faced

  • Designing realistic emotional and contextual responses that feel natural.
  • Balancing medical accuracy with conversational fluidity.
  • Handling complex branching dialogues without breaking immersion.
  • Integrating voice and text modalities while maintaining low latency.

What We Learned

We learned how powerful agentic AI systems can be when fine-tuned for education. Building Clyra showed us the potential for AI as a scalable medical tutor — not to replace clinical experience, but to amplify it.

What’s Next

  • Expanding the case library across multiple specialties.
  • Mobile App

Built With

Share this project:

Updates