Back to work
Independent build2026In active development

Medicine Helper

A camera-first medicine label assistant designed to turn an unfamiliar medicine pack into a clearer, safer next step.

HealthcareVision AISafety by design
Business problem

People often have medicines at home without a clear understanding of what they are for, their risks, or what information deserves attention before use.

Key decision

I designed the experience around camera-first identification, structured medicine facts, explicit safety scoring and constrained guidance rather than an open-ended medical chatbot.

What changed

The concept progressed from a Custom GPT into a native mobile product flow with label scanning, curated medicine information, safety cues and alternatives.

Takeaway

In high-stakes products, the design of uncertainty, guardrails and escalation is part of the product itself, not a disclaimer added at the end.

Want the implementation details?

Open the full architecture, modeling, validation and lessons.