DinoSync

Inspiration

As a prior healthcare professional who oversaw 100k+ patients, I witnessed brilliant clinicians drowning in documentation while patients waited. I saw talented doctors spending more time typing than treating, nurses buried in paperwork instead of providing care. The irony was painful, technology that promised to help was actually pulling clinicians away from what they do best. Healthcare doesn't need another robot — it needs a true collaborator.

What it does

DinoSync is an AI platform that collaborates with clinicians rather than replacing them. It listens to natural clinical conversations, learns individual preferences, and adapts in real-time to provide immediate, accurate documentation and personalized care suggestions. Our system integrates seamlessly into existing workflows, respects clinical decision-making, and continuously improves through collaboration with healthcare providers.

How we built it

We built DinoSync using React 19.1.0 with TypeScript for a type-safe, modern frontend experience. Our component library leverages Material-UI v7 with Emotion for CSS-in-JS styling. The core architecture uses React Hooks for state management and integrates Whisper AI for voice-to-text transcription and Mistral AI LLM for clinical text generation and SOAP note creation. We developed a custom adaptive learning service in TypeScript that learns clinician preferences. Audio capture utilizes the Web Audio API for real-time recording and visualization, while axios handles API communications.

Challenges we ran into

Privacy & Security: Building HIPAA-compliant infrastructure from day one required extensive security protocols and data encryption. Clinical Accuracy: Developing AI that understands medical nuance demanded specialized training on verified clinical datasets. Integration: Creating APIs that work with diverse EHR systems meant navigating complex healthcare IT environments. Trust: Designing transparent AI that explains its reasoning to skeptical clinicians required careful UX design and extensive clinical feedback.

Accomplishments that we're proud of

We've built a production-ready architecture with seamless mock-to-API integration capabilities. Our component-based design with Material-UI creates an intuitive clinician experience, while our custom adaptive learning service successfully captures and applies individual preferences. The real-time voice processing with Web Audio API provides smooth, responsive audio capture, and our TypeScript implementation ensures maintainable, scalable code ready for hospital deployments.

What we learned

The most crucial insight: clinicians don't want to be replaced — they want to be empowered. When AI respects human expertise and enhances clinical judgment rather than overriding it, magic happens. Documentation becomes effortless, care becomes more personalized, and clinicians rediscover their passion for healthcare. Collaboration beats automation every time.

What's next for DinoSync

We're transitioning from mock implementations to live Whisper and Mistral API integrations, expanding our adaptive learning algorithms, and preparing for HIPAA-compliant deployments. Our roadmap includes advanced clinical modules, EHR integrations using our scalable React architecture, and mobile companion apps. We're also exploring additional AI models that can plug into our flexible TypeScript service layer for specialized medical tasks.

Built With

• audio
• axios
• material
• mistral
• openai
• react
• whisper