Inspiration
We wanted to create an AI-powered solution which efficiently connects doctors with patients, bridging gaps in communication and documentation fatigue. In a world where doctors spend more time on paperwork than patients, and patients struggle to articulate complex health trends, we saw an opportunity to use Multimodal AI to create a "living" health record that speaks the language of both sides.
What it does
Our app is a dual-sided platform that synchronizes the patient journey with the clinical workflow:
For Patients: An intuitive health diary where users log symptoms, food, and mood via voice or text. It allows for "Visual Logging," where AI analyzes photos of symptoms (like rashes or wounds) to track progress.
For Doctors: A high-efficiency dashboard that transforms messy patient data into structured SOAP notes. It also functions as a Clinical Note Cleaner, taking a doctor's raw, unstructured voice dictations and instantly distilling them into professional medical documentation.
How we built it
We built a sophisticated multimodal pipeline using the best of Azure and Flutter:
Frontend: Built with FlutterFlow for a rapid, responsive mobile experience.
Backend & Data: Integrated Supabase (PostgreSQL) for real-time data synchronization and secure patient record storage.
The AI Brain: Custom prompted GPT-4.1-mini: Prompts precisely instructing the AI how to summarize doctor-patient dialogues.
Azure OpenAI: Orchestrates the final SOAP note synthesis, combining text summaries with visual descriptors.
Azure AI Vision: (Possible extension) Processes patient imagery to provide objective descriptions of visible symptoms.
Challenges we ran into
Privacy: Balancing the need for cloud-based AI processing with strict doctor-patient confidentiality.
Technical Roadblocks: Navigating the limitations of free-tier cloud services under a 12-hour deadline forced us to pivot our backend architecture multiple times mid-sprint to achieve a working demo.
Log in or sign up for Devpost to join the conversation.