We didn't start with a feature list. We started by asking patients why their portal made them anxious, and clinicians why eight minutes never felt like enough — then built backward from there.
Most people managing type 2 diabetes are doing the hard part alone: testing, logging, remembering what the doctor said last time, and trying to make sense of lab numbers that arrive with no context. Then they get eight minutes with a clinician who's seeing the same numbers from the opposite side — buried in billing-driven software that wasn't built for either of them.
BloodSugar started as a simple question: what if the same health data could be shown two honest ways — calm and plain for the person living with it, complete and clinical for the person treating it — without either side losing trust in the numbers?
We built it talking to patients and clinicians, not just designing for them. Every feature here — from where a lab value came from, to how an AI suggestion is worded — exists because someone told us where the last tool let them down.
Every value carries where it came from — a lab, a scan, a synced device, or typed by hand — so nothing is taken on faith.
The patient screen and the clinician report read the same dataset differently — they never disagree.
Our AI flags patterns worth a conversation. It never tells you what's wrong, or what to take — that's still your doctor's job.
This is an active UX prototype. Every stage below changed something real after talking to the people who'd actually use it.
Before writing a line of code, we looked at where EHR charting, patient portals, and data visualization actually break down for both sides of the exam room.
Built the core idea: a calm patient dashboard and a dense clinician report, generated from the same records instead of two disconnected tools.
A course supervisor's review pushed us to build the parse-and-verify step before trusting any scanned value, and to show exactly where every number came from.
Reminders, a prescription vault, accessibility settings, and Bluetooth meter sync all came directly from interviews with people who'd actually use this daily.
From categorizing overwhelming charts by health domain, to turning off emoji by default — the product keeps changing because the feedback hasn't stopped.
This prototype is still evolving — every screen here came from someone's feedback.