mobile version
Add as many family members as needed, and directly send them reports
All medicine and their status, as simple as possible
Family summary, more clinical and precis
Simple recording/pausing with live transcript
Patient Summary, Necessary info, and Message forwarding

ScribeDoc 🩺

Your Personal Advocate in the Doctor's Office 🏆 2nd Place Winner – Healthy Aging Hackathon

ScribeDoc is an ambient AI companion that bridges the critical communication gap between older adults, their physicians, and their families. By acting as a secure, invisible scribe, it transforms complex 15-minute medical visits into simple, actionable summaries without disrupting the doctor-patient relationship.

💡 The Inspiration

In a short clinic window, patients are expected to process complex clinical changes and advocate for themselves, often while managing the baseline anxiety of being in a medical setting. Unless a family member is present to take furious notes, vital instructions can easily vanish the moment the patient leaves the room. We built ScribeDoc to ensure no patient is left confused and no caregiver is left out of the loop.

✨ Key Features

Dual-Audience Translation: The AI translates messy, jargon-heavy dialogue into two distinct views: a highly simplified, encouraging checklist for the patient, and a precise clinical overview for the family.
Dynamic Medication Tracking: Automatically parses the conversation for medication changes, visually categorizing them (New, Stopped, Continued, Adjusted) so the patient immediately knows their current regimen.
Native Accessibility: Features a 125% scaled UI for older eyes and a built-in Text-to-Speech (TTS) engine that reads out generated follow-up questions, allowing patients to simply hold up their device to ask the doctor clarifying questions.
Frictionless Continuity of Care: With one tap, the structured summary, raw audio, and transcript can be emailed securely to secondary caregivers.
Multilingual Support: Because the engine processes raw audio, it perfectly extracts clinical details regardless of the spoken language (successfully tested in English, Urdu, Chinese, and Arabic).

🛡️ Our Philosophy: AI as a Supplement

ScribeDoc is a supportive communication bridge, not a robotic doctor. To ensure patient safety, our AI operates under strict clinical constraints:

Zero Medical Recommendations: ScribeDoc never oversteps its role. It does not diagnose and will absolutely never suggest a treatment or lifestyle change that was not explicitly stated by the physician.
Immutable Source of Truth: Every AI-generated summary is anchored to the original raw audio and transcript, ensuring patients can always verify instructions directly from the doctor's mouth.
Safety Flagging: The system acts as a safeguard, warning the user to follow up with their provider immediately if severe warning signs or dangerous combinations are detected.

🛠️ How We Built It & UX Guardrails

Tech Stack: Rapid full-stack web application built with Next.js and Tailwind CSS, utilizing native browser APIs to keep the tool lightweight and fast.
AI Engine: Powered by Google's multimodal Gemini 2.5 Flash. By passing raw audio directly to the LLM, we bypassed traditional transcription latency for near-instant analysis.
Fail-Safe UX: In high-stress environments, accidental screen taps happen. Tying the "End Visit" action to an immediate stop risked losing vital data. We architected a custom confirmation overlay that keeps the microphone actively capturing in the background, ensuring zero data loss during user error.

🚀 The Road Ahead

To transition ScribeDoc from a solid prototype into a real-world healthcare application, our next phases of development include:

Secure Cloud Infrastructure: Moving to a fully HIPAA-compliant backend so remote family members can securely log into a real-time patient dashboard.
Native Mobile Ecosystem: Dedicated iOS and Android applications for deeper integration with native accessibility tools.
EHR Integration: Adopting the FHIR standard to automatically cross-reference ScribeDoc's extracted medication lists with a patient's official electronic health records to alert doctors of potential discrepancies.