Building PGIS: A Vibe-Coded Performance Glycemic Intelligence System

Why I’m tracking blood sugar like a developer — and training like an athlete

As a physician who codes, I’ve learned that the most durable software projects rarely begin as startups. They begin as personal pain points.

Over the past year, I’ve been quietly building something I now call the Performance Glycemic Intelligence System (PGIS) — a vibe-coded, personal web and mobile application designed to help me understand the relationship between training load, recovery, and glucose behavior in real time.

This is not a commercial product.
This is a side project with purpose.

The motivation: aging, diabetes, and endurance

I’m a long-standing type 2 diabetic, approaching 60, and currently training toward a Thanksgiving 2026 half marathon. Those three facts intersect in a way most traditional health apps simply don’t address:

• Blood glucose is not static

• Training stress is cumulative

• Aging changes recovery kinetics

As clinicians, we understand these concepts intuitively — yet most consumer tools flatten them into dashboards that lack narrative, context, and longitudinal insight.

PGIS was born out of that gap.

What PGIS actually is

PGIS is a personal intelligence layer, not just a tracker.

At its core, it ingests:

• Continuous glucose monitor (CGM) data

• Training sessions (running, walking, strength, mobility)

• Subjective recovery and inflammation scores

• Stream-of-consciousness notes recorded immediately after sessions

These raw inputs are then transformed into structured, clinician-grade summaries using markdown schemas and narrative synthesis — suitable for long-term trend analysis, reflection, and decision-making.

Think of it as:

A personal metabolic “training log,” written the way a physician would want to read it.

Vibe coding, intentionally

This project was vibe coded — deliberately.

I didn’t start with:

• a product roadmap

• a monetization plan

• or a polished UI

I started with:

• my Garmin screenshots

• my CGM graphs

• and my own clinical curiosity

PGIS evolved organically because I care deeply about the outcome. That’s a lesson I’ve learned repeatedly while coding:
interest sustains iteration.

The tech stack (light, flexible, personal)

PGIS currently lives as:

• A private GitHub-hosted system

• Markdown-driven summaries

• AI-assisted synthesis using NotebookLM-style workflows

• A Next.js dashboard layer for review and annotation

The architecture is intentionally simple — optimized for clarity, not scale.

The repository is available on GitHub, for those interested in exploring or adapting the approach for their own use.

How this fits alongside CodeCraftMD

Some readers have asked how PGIS relates to CodeCraftMD.

They are complementary — but distinct.

While I’m actively hardening CodeCraftMD toward HIPAA-compliant workflows, PGIS remains a personal R&D lab:

• No patient data

• No clinical deployment

• No commercialization pressure

That freedom is precisely what allows experimentation.

Side projects like PGIS sharpen skills that ultimately make larger platforms stronger.

An invitation to other clinicians

If you’re a clinician who:

• tracks glucose, blood pressure, sleep, or training

• wants narrative intelligence instead of raw numbers

• or is curious about building personal health tooling

You don’t need permission.
You need a problem that matters to you.

If you’d like:

• guidance on structuring a similar system

• help designing a clinician-friendly schema

• or consultation on personal health analytics

I’m happy to connect.

PGIS documentation & printable summaries

Below are reserved sections for printable PDF outputs generated from PGIS summaries — useful for reflection, coaching discussions, or long-term archival:

The development of the Performance Glycemic Intelligence System (PGIS) exemplifies the intersection of personal initiative and impactful problem-solving. This innovative platform empowers users to gain insights into their metabolic responses, fostering a deeper understanding of health dynamics. By leveraging advanced coding techniques, PGIS transcends traditional approaches, offering a comprehensive view of the body as an interconnected system. In essence, the most significant advancements often emerge from projects driven by personal relevance and necessity.

PGIS_Performance_Metabolic_Resilience_Case_Study (5)Download

Run_Strategy_Metabolic_ControlDownload

(Generated via NotebookLM-assisted synthesis and markdown-to-PDF workflows.)

Final thought

Doctors are trained to interpret complex systems.

Coding simply gives us another lens.

PGIS isn’t about optimizing glucose alone —
it’s about understanding the body as a system under load, over time.

And sometimes, the most meaningful software we build
is the kind we build for ourselves.

— Dr. Chukwuma Onyeije
Doctors Who Code