MedLens is a local-first medical tracking app for patients and family caregivers. When a loved one is hospitalized — especially for long-term care such as cancer treatment — the information that actually matters is scattered across verbal updates, photographed whiteboards, IV bag labels, lab printouts, and handwritten notes. MedLens turns that bedside chaos into a structured, longitudinal record: point your camera at an artifact, on-device OCR extracts the data, you confirm or correct it, and trends accumulate over weeks of treatment. Everything stays on the device. It is built with React Native and Expo, ships to the Apple App Store and Google Play, and is the patient-side half of a deliberately firewalled two-app patient-safety stack.
Quick Links
- Architecture — Expo Router app shell, layered
app/+src/structure, state model - Data Model — the SQLite schema, the medications-vs-logs and panels-vs-results hierarchies, patient-delete invariants
- OCR Extraction Pipeline — the three-tier capture flow and the privacy ladder behind it
- Trends and Charts — context-aware coloring, where “up” is not always good
- Privacy and Safety — local-first storage, the SaMD firewall, notification PII rules
- Developer Guide — running a dev build, testing, and where things live
Key Facts
- What it is: a patient-side capture journal — photograph hospital artifacts, extract structured medical data, track trends over time, and share a sealed snapshot with a clinician on demand.
- What it is not: not a diagnosis tool, not a patient-portal replacement, not a cloud service. Clinical interpretation lives outside MedLens by design. See the SaMD firewall.
- Stack: React Native / Expo SDK 55, Expo Router (file-based navigation), expo-sqlite, on-device ML Kit OCR, Victory Native charts, TypeScript strict mode.
- Storage: local-first — all records in SQLite plus a structured photo directory on the device filesystem. No server holds your medical data.
- Status: substantially feature-complete across all five design phases (capture, OCR, trends, AI enhancement, polish); approaching first App Store submission.
Source of Truth
The canonical design, schema, and OCR pipeline specification live in the main MedLens repository at github.com/blamechris/medlens. This wiki mirrors that design at a portfolio level; the repository is private and remains authoritative.