Sovereign Health Stack — Final Scoring & Ranking

A sport tracking app explicitly framed around privacy: the F‑Droid listing highlights no Internet access, no ads, and an Apache‑2.0 license. The project also maintains an official website describing privacy-first sharing controls.

A lightweight, open-source DICOM server positioned as a local “mini‑PACS” with a web interface, REST API, and a standalone architecture. The Orthanc Book provides explicit guidance on securing DICOM flows and REST API exposure.

Licensing nuance: Orthanc’s core server is GPLv3+; Orthanc’s own download page notes that some plugins/viewers may be AGPLv3+.

A BSD‑licensed, openly developed imaging interaction toolkit and application framework built on top of ITK and VTK. Strong local/offline posture with a high composability score, balanced by a real operational/engineering footprint.

A cross-platform toolkit for N-dimensional image processing, segmentation, and registration. As a library it has no inherent network surface: privacy properties are determined by the host application and deployment.

A foundational scientific visualization toolkit used globally across research and industry. Like ITK, it is local by design; operational cost is in build/tooling rather than in data-leak surfaces.

The F‑Droid listing states: no ads, no trackers, no data collection, and it works even if Internet access is blocked. Narrow scope (med reminders) but extremely clean from a privacy/offline standpoint.

The F‑Droid listing states it has no ads and requests no unnecessary permissions; location permission is only used to discover Bluetooth scales and can be revoked after pairing. It does not require an account.

Note: there is an optional openScale sync add-on which can synchronize to external services (e.g., Health Connect / MQTT) — that is optional and not required for local operation.

F‑Droid describes MedTimer as fully offline and privacy-focused, storing all data on-device with no Internet connection required. MIT license, no ads, and local backups/exports.

A free and open-source platform for medical image data analysis, with a BSD-style license. Core usage is local/offline, but telemetry can be introduced via extensions (notably SlicerTelemetry), and community guidance emphasizes user consent when collecting usage data.

A multifunctional, modular DICOM viewer (standalone and web-based). The documentation states it is dual-licensed (EPL‑2.0 or Apache‑2.0), providing flexibility for different deployment needs. Cloud integrations exist, but the default posture remains broadly on‑prem and viewer-centric.

A GNU Project electronic medical record emphasizing longitudinal care and patient privacy. Strong licensing and self-hostability, with the main deductions coming from operational complexity and clinic-style workflow assumptions.

A full hospital and health information system with strong copyleft licensing (GPL v3 or later) and explicit self-hosting posture. The main penalties come from operational burden and the fact that workflows are shaped around institutional standards and large deployments.

An enterprise-focused open source LIMS that covers laboratory workflows end-to-end. The core repo states it is GPLv2 and derived from BIKA LIMS. Strong sovereignty posture via self-hosting, with deductions reflecting operational complexity and admin burden.

A zero-footprint web DICOM viewer (progressive web app) designed for DICOMweb archives. Architecturally it can be deployed against self-hosted DICOMweb sources; however, its official documentation includes a dedicated Google Cloud Healthcare integration path, creating strong ecosystem gravity toward big-cloud deployments.

Self-hosted Personal Health Record (PHR) manager with a clear “data stays local” posture. The repository README explicitly states that the OnPrem app does not integrate with EHRs directly; it supports manual entry and importing FHIR bundles exported through other means.

External-service caveat: the Fasten OnPrem releases include a CCDA feature note “powered by Health Samurai” and state that the converter is not open source and runs on Fasten Health infrastructure (no data sent back to Health Samurai during conversion).

A web-based CGM dashboard for remote viewing of glucose data. The official documentation describes installing the Nightscout cloud application using hosting services from various cloud providers, and “new user” docs enumerate popular vendor platforms.

Score deductions come primarily from: (1) reliance on proprietary CGM hardware ecosystems, and (2) cloud-forward deployment patterns.

An open source artificial pancreas system running on Android. Documentation states that use requires three compatible devices: an Android phone, a CGM, and an approved insulin pump. APK distribution is constrained by medical device regulations, so the docs emphasize building the app rather than downloading.

Score deductions primarily reflect deep reliance on proprietary CGM/pump ecosystems and regulated supply chains, despite strong open-source posture.