AI Medical Visit Companion

HealthiHer is a responsive web application that helps users feel more prepared and supported through the full medical appointment lifecycle — before, during, and after a doctor visit.

Instead of a chatbot UI, HealthiHer uses a journal-based interface with structured cards, checklists, and visit timelines to reduce cognitive load and make medical information easier to understand and revisit later.

• Visit metadata (doctor name, reason, date/time)
• Symptom + goals notes
• AI-generated Suggested Questions
• Save/star questions + add custom questions

• In-browser recording (Start/Stop)
• Recording timer + active indicator
• Quick access to prepared questions (tap to mark as asked)

• Plain-language visit summary (bullet points)
• Follow-up & next steps checklist
• Medical terms explained (jargon -> simple language)
• Suggested follow-up questions for future visits
• Bias detection score (flags dismissive language patterns)

• Quick slider/rating prompts (heard, clarity, comfort)
• Optional checklist + reflection notes
• Saved to the visit record

Doctor visits can be stressful, rushed, and easy to forget and patients often leave with unanswered questions or unclear next steps. HealthiHer is designed as a visit memory system + private health journal, helping users advocate for themselves with more clarity and confidence.

• React 19
• TypeScript
• Vite
• Tailwind CSS
• Framer Motion

• OpenAI Whisper (speech-to-text transcription)
• OpenAI GPT-4.1-mini (question generation, summaries, next steps, jargon)

• localStorage

• Vite Dev Server Middleware

High-level flow:

1. User records audio in the browser
2. Audio is sent to Whisper for transcription
3. Transcript + visit context is analyzed by GPT-4.1-mini
4. App renders structured outputs (summary bullets, next steps, questions, terms)
5. Results are saved to localStorage as a visit record

This repo is organized around screens/pages that match the user flow:

• Authentication
• Journal timeline (home)
• Prepare for appointment
• Appointment assistant (recording mode)
• Visit summary & insights
• Reflection

git clone https://github.com/nssim516/hackviolet2026.git
cd hackviolet2026

npm install

Create a .env file in the project root:

VITE_OPENAI_API_KEY=your_key_here

Note: This project uses OpenAI via middleware routes. Do not commit your key.

npm run dev

http://localhost:3000/

• Secure authentication + database-backed visit history
• Improved real-time follow-up question batching during appointment mode
• Export/share formats (PDF, share link)
• Interactive symptom map / body diagram input
• More personalization by age, specialty, and medical context