Healthcare

Anchor Personal Health Platform

A private, local-first health platform for navigating a serious diagnosis. Anchor pulls records, labs, medications, nutrition, and the latest treatment research into one place and separates signal from noise so the patient can make confident, informed choices and advocate for themselves. Its tagline says it plainly: steady through the journey.

The Challenge

Facing a life-threatening diagnosis, a patient is buried in fragmented data and conflicting advice. Lab PDFs, scanned pathology, imaging, insurance statements, provider bills, and medication schedules arrive from a dozen places that never talk to each other. On top of that comes a flood of treatment claims that range from rigorous science to profit-driven snake oil, with no easy way to tell which is which. No consumer tool ties it all together, speaks plain language instead of clinical jargon, shows who profits from each recommendation, or treats diet, movement, mind, and faith as part of healing rather than an afterthought. And because the data is deeply personal, it must never leave the patient's control.

The Solution

I designed and built Anchor with Claude Code as a local-first platform that runs on the patient's own machine. A Python and FastAPI backend with a React and TypeScript frontend keeps every record on-device, protected with field-level encryption, multi-factor login, and an append-only audit trail, so medical data never goes to a third-party cloud. It brings the whole picture into one place: Health Records that ingest labs, vitals, imaging, genetics, and pathology (with CCDA and FHIR import, standardized lab coding, and OCR for scanned documents); a conditions and episodes timeline; visits with appointment and visit-prep briefs; nutrition with meal planning, grocery lists, and supplement-interaction safety checks; a medication calendar; and a Body and Spirit space for movement, sleep, prayer, and journaling. A proactive research engine scans PubMed, ClinicalTrials.gov, and integrative sources, labels every source by type so non-commercial evidence sorts first and anything with a financial incentive is marked, and sends a weekly digest. A billing module cross-references insurance statements against provider bills to catch overcharges. A Personal Compass of preferences and a tone slider tailor what each screen surfaces and how it speaks. On-device AI handles the extraction and synthesis so nothing sensitive is sent away.

The Impact

Anchor turns a scattered, overwhelming diagnosis into one calm, private workspace. Records that arrived as PDFs and scans become structured, searchable data. Treatment options arrive pre-sorted by credibility with the financial incentives labeled in the open. Medications, nutrition, and appointments live on one calendar, and body-and-spirit practices sit alongside clinical care instead of in a separate wellness silo. Every recommendation surfaces the trade-off in plain words and leaves the decision to the patient. Because it is local-first, none of it leaves the patient's machine. It is a working example of an AI-augmented product with real depth, medical document parsing, encryption, and proactive research, built by one person with modern tooling.