Patient Monitor

PatientMonitor: Smarter Health Insights for Doctors About the Project

PatientMonitor is a real-time health tracking platform built to support family doctors managing multiple patients — especially seniors and those with chronic illnesses — through a clean, connected interface.

Patients log their daily blood pressure readings, mood, exercise, and food intake via a mobile app. Doctors access a powerful dashboard that visualizes trends and allows them to interact with an AI assistant. This assistant, powered by a Retrieval-Augmented Generation (RAG) pipeline, leverages the patient’s history, local medical data, and public health guidelines to provide intelligent, context-aware insights—helping doctors diagnose more efficiently and accurately using the patient’s health data.

What Inspired Us

Our inspiration came from a real-world gap: many seniors record their blood pressure manually or sporadically, and family doctors are often overwhelmed with fragmented information.

We envisioned a simple tool that could:

Help patients consistently log BP readings.

Help doctors focus on patients, not paperwork.

Use AI not just for automation — but for insight.

How We Built It

We combined powerful, no-code/low-code tools with solid data architecture:

Frontend: Built using Lovable.dev for rapid UI prototyping.

Backend: Supabase handled patient data, readings, and doctor assignments.

AI Assistant: We used Vectorize.io to build a RAG pipeline. The chatbot retrieves insights from patient summaries, Ministry guidelines, and pharmacy/lab directories.

We also generated synthetic patient histories and 14-day BP datasets using Python.

Challenges We Faced

Managing data realism: We needed synthetic but medically plausible data — so we crafted custom patient profiles with diverse conditions and daily BP fluctuations.

RAG tuning: Integrating Vectorize’s retrieval system and getting it to respond naturally required trial and error, especially chunk size tuning and metadata filters.

Lovable integration: While Lovable made it easy to build fast, connecting dynamic APIs and formatting chatbot responses required manual intervention.

Time: Balancing real-time data, AI chatbot, and frontend UI — all during a hackathon — was an intense but rewarding challenge.

🚀 What We Learned

The value of building for real users — this tool could genuinely help doctors and patients.

How to combine structured data (Supabase) with unstructured knowledge (Vectorize RAG) in one product.

That no-code tools like Lovable can be incredibly powerful when used alongside code, not instead of it.

“PatientMonitor bridges the gap between patient effort and medical insight — empowering both patients and the people who care for them.”

Built With

• lovable
• supabase