Custom supplements based on a blood test

Most supplement recommendation products hand everyone broadly the same stack, adjusted by a questionnaire. Custom supplements based on a blood test are a different product only if the blood test actually drives the result: the person’s profile decides which panel to order, and the numbers that come back decide the protocol. Anything else is an expensive guess with a lab report attached.

Perfect Lyfe was built end to end on that order, as both a web app and a mobile app. A person creates a profile with their basics and a goal, a blood panel is recommended and a lab order generated, they get the bloodwork done and upload the result, and biomarkers are extracted from the uploaded PDF, with manual entry as a fallback. A rules engine then turns those biomarkers into a personalised daily protocol.

The fallback is the part that decided whether any of it could be trusted. Lab reports arrive in whatever layout the laboratory prefers, and a value read into the wrong field is worse than a missing one, because everything downstream trusts it. Letting a person confirm or enter the numbers themselves keeps the rules engine working from values someone has looked at, rather than from whatever the extraction happened to produce.

Before any of that, sign-up and onboarding had to carry a person from a first visit to a recommended panel. Around that core sat the parts that make a protocol something people actually follow: a dashboard with a score, streaks and trends, daily tracking with reminders, and a warning flow for biomarkers that came back low. It got unusually far for a project in this category, all the way to live.

Then it was taken down, and the code no longer exists. The case study documents what was built rather than pretending it still runs, and records that it reached live before it was removed.