Inspiration Healthcare is overwhelmed. Patients wait weeks for rushed consultations, and doctors are drowning in admin work. The result? Delayed care, frustrated patients, and overburdened physicians. Our AI-powered health kiosk changes that—offering real-time assessments, tracking key health markers, and structuring patient data before a doctor visit. It’s not about replacing doctors but making every consultation count.
What It Does Our kiosk acts as an AI-driven medical assistant, guiding patients through voice-based consultations, analyzing real-time health data from wearables and medical devices, and even assessing visual symptoms via integrated cameras. It generates differential diagnoses, treatment plans, and specialist referrals—bringing efficiency and accessibility to healthcare.
How We Built It Frontend: Next.js, WebRTC (real-time voice), TailwindCSS Backend: Next.js API routes, Prisma ORM (PostgreSQL), BullMQ (job processing) AI & Processing: GPT-4o, Vision Analysis, multimodal medical reasoning Data & Storage: AWS S3, Google Fit API (biometric tracking) Authentication & Security: Clerk (user sessions), OAuth (Google Fit), presigned URLs (secure document access) Challenges Faced WebRTC & AI Latency: Ensuring seamless voice interaction
Accomplishments ✅ Fully autonomous AI consultations with multimodal reasoning ✅ Seamless integration with medical devices & biometric tracking ✅ Proactive checking for conditions such as heart attacks/stroke ✅ Scalable, real-time architecture handling multiple patient sessions ✅ Google Fit integration for real-time health data analysis from wearables
What We Learned Modular Design Scales Better – Next.js + TRPC keeps components flexible andBullMQ reduced response lag significantly
Future Work 🚀 More Sensors – Blood glucose, ECG, and expanded biomarker detection 🚀 FHIR & HL7 Integration – Syncing with hospital EHRs 🚀 Multi-Language Support – Expanding accessibility
Built With
- nextjs
Log in or sign up for Devpost to join the conversation.