Private medical archive

Readable records, original documents, and corrections in one place.

Medikeep turns scattered lab PDFs, scans, and prescriptions into a local-first medical archive. It preserves the original source file, extracts structured results for fast reading, and lets you correct mistakes directly instead of editing a wall of OCR text.

Open Medikeep on GitHub How it works

Offline-first Import, OCR, parse, store, and review without shipping health data to the cloud by default.

Source-backed Every saved record keeps the original file for verification instead of hiding the source away.

Editable Correct rows, values, units, and ranges directly inside the structured report view.

Medikeep

Record detail

Complete Health Check — 26 Oct 2025

Original PDF preserved · Structured results extracted · Reviewable source

26 Oct 2025 Blood sugar Cholesterol

Measurement	Value	Unit	Range
HbA1c	5.7	%	4.0 – 5.6
Estimated Average Glucose	117	mg/dL	—
HDL Cholesterol	37.6	mg/dL	> 40

Extraction note: results remain linked to the original source because medical OCR and PDF parsing still need human verification.

What Medikeep solves

Medical records are usually stored as clutter, not knowledge.

Families accumulate prescriptions, test reports, and scans across years. Most files are visually readable but structurally useless, especially when they come from different labs and different layouts. Medikeep keeps the file, extracts what matters, and still lets the user check the source.

01 · Import

PDFs, scans, images, and text

Bring in machine-generated PDFs, scanned reports, prescription images, or plain text and keep the original file archived inside the app.

02 · Structure

Flexible results, not fixed schemas

Medikeep stores sections and result rows so different labs and different panels do not break the model every time a document layout changes.

03 · Correct

Edit the structured output directly

Fix OCR and parsing mistakes at the row level instead of scrolling through a massive block of extracted text that no one wants to edit.

How it works

Built for the actual mess of medical documents.

The workflow is designed around inconsistent lab layouts, mixed-quality scans, and the fact that health data should remain reviewable and user-controlled.

Archive the original

The source file is copied into the app’s private archive so the document is still available for verification later.

Extract text locally

PDF text extraction is preferred when available; OCR is used for scans and images when the document lacks a reliable text layer.

Map results into sections

Medical terms are grouped into consumer-readable areas like Blood sugar, Cholesterol, Thyroid, and Blood count.

Review and correct

The user can compare the structured output against the original file and fix specific rows, dates, units, or ranges directly.

Why the product direction matters

Privacy first. Trust first.

A medical archive should not feel like a generic OCR demo.

Medikeep is being built as a real personal records layer: original document preview, structured rows, editability, and an offline-first default for sensitive data.

Local-first import and processing by default
Original source file preserved for audit and comparison
Structured results designed for long-term readability
Parser iteration backed by real report corpora, not only synthetic examples

Medikeep

Private records should stay readable, reviewable, and yours.

Medikeep is currently being developed as an Android-first medical archive with structured parsing, editable results, and source-preserving imports.

Open GitHub repo See more Upthink projects