Projekt-Homepage:
https://help-cure-als.org/projects/alsnews/
An open-source platform for turning medical and clinical articles (studies, papers, registry entries, newsletters, press releases) into clear, layperson-friendly explanations — without compromising scientific accuracy.
The focus is on structured summaries, transparent sources, and a stable, long-term process that can be operated by both technical and non-technical contributors.
-
Directus
- Content backend
- Editorial UI (creation, review, publishing)
- Roles, policies, permissions
- Triggers (Flows → Webhooks)
-
Automation Backend (Express + TypeScript)
- Stable business logic outside Directus
- Scraping and metadata extraction
- Deduplication (DOI / PMID / Trial ID / URL / content hash)
- AI-based summarization and translation
- Update detection and linking (e.g. study or medication updates)
-
PostgreSQL
- Single source of truth
- Includes Directus system data (users, roles, flows, layouts)
-
Optional uploads
- PDFs, images, supplementary files via Directus
- Docker
- Docker Compose
From the repository root:
cp .env.example .env
docker compose up -d./scripts/seed.shAfter startup:
- Directus UI: http://localhost:8055
- Automation backend: http://localhost:3010
The following account is available for development and testing purposes:
- Email:
[email protected] - Password:
admin1234 - Role: Administrator (full access)
⚠️ Note:
These account is intended for local development and testing only.
Make sure to change or remove these credentials in any production environment.
Articles use Directus Translations:
- one canonical article record
- language-specific text stored separately
- ratings, sources, and update logic remain centralized
All operational scripts (backup, restore, cleanup) are located in:
scripts/
Each folder contains its own README with concrete instructions.
Contributions are welcome in the following areas:
- Source integrations (registries, journals, feeds)
- Scraping and extraction logic
- Editorial review and language clarity
- Translations
- UX improvements in Directus
- Prompt design and evaluation
Please open issues with:
- source URL or identifier
- expected vs. actual behavior
- anonymized example data if possible
License: MIT
This project does not provide medical advice. All content is informational and must be interpreted in consultation with qualified medical professionals.
All editorial content, summaries, and explanatory texts are licensed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0), unless stated otherwise.