AZYGOS: AI-Assisted OSCE Exam Simulator

Medical students in Switzerland face one of the toughest checkpoints of their training: the OSCE, a practical clinical exam where every history taken, lab ordered, and diagnosis reached is scored against strict criteria. Realistic practice for it is scarce, costly, and hard to schedule.

AZYGOS answers this with a browser-based OSCE simulation platform. Students work through interactive clinical cases, either alone or in live multiplayer sessions that rotate them through candidate, patient, and examiner roles. Each case is generated and adapts to the student, and an automated scoring engine grades performance against the exam’s own criteria, returning structured feedback.

Built with Next.js and TypeScript on a Supabase backend (Postgres, Auth, Realtime), the platform pairs clinically grounded content with a fast, collaborative interface. It is fully trilingual (German, French, Italian) and hosted entirely in Switzerland to meet medical data-residency requirements.

Designed and built end to end through agentic coding, from the case pipeline and real-time multiplayer to the scoring engine and trilingual UI.