CareLog is a specialized digital solution engineered to alleviate the administrative burden within the Hungarian social care system. It addresses the critical challenge of time-consuming manual documentation, which often detracts from direct patient care.
By leveraging Voice-to-Text technology and Large Language Models (LLMs), CareLog enables care workers to dictate logs and receive intelligent, regulation-compliant advisory support entirely in Hungarian.
- Cross-Platform Mobile App: Built with Ionic and React for a seamless experience on both iOS and Android.
- Hungarian Voice-First Data Entry: Care workers can dictate logs directly into the mobile app in Hungarian, saving time and reducing typing effort.
- AI Knowledge Support (RAG): A Retrieval-Augmented Generation module helps navigate complex care protocols and documents, providing instant, context-aware information to support decision-making.
- Dockerized Deployment: Fully containerized environment for easy setup and reproducibility.
- Runtime: Node.js
- Framework: Express.js
- API Spec: TSOA (OpenAPI/Swagger)
- Database: PostgreSQL + Prisma ORM
- Language: TypeScript
- Language: Python
- LLM: Google Gemini API
- Embeddings: Vector-based document retrieval
The system follows a modern client-server architecture with a specialized microservice for AI processing. The following components are used:
- Frontend: React + Ionic
- Backend: Node.js + Express.js
- Databases: PostgreSQL + Elasticsearch (used as a vector database)
- Worker node for async tasks: Node.js + Express.js
- Redis for message queue
- AI and pre-processing RAG module: Python + Google Gemini API
- Docker compose
- A valid Google Gemini API Key
git clone https://github.com/your-username/carelog.git
cd carelogThe system requires simple configuration variables.
- Navigate to
backend/and create a.envfile. - Add your Google API Key:
GOOGLE_API_KEY=your_api_key_here
See INSTALL.md for full configuration details.
We provide a specialized script to launch the entire stack in one go.
# On Linux/Mac
./run.sh
# On Windows
./run.batAccess the applications at:
- Frontend: http://localhost
- Backend API: http://localhost:8080
cd backend
npm install
npm run devcd frontend
npm install
npm run devFor detailed contribution guidelines, please see CONTRIBUTING.md.
This project is licensed under the MIT License - see the LICENSE file for details.
Created by Roland Csősz

