simplazbluda/CDS-MAS

Multi-agent AI system for clinical triage and decision support, designed for offline-first deployment in low-resource healthcare environments. CDC-MAS - Clinical Decision Support- Multi-Agent System 1. Project Overview CDS-MAS is an AI-powered Clinical Decision Support System (CDSS) designed for low-resource healthcare settings. The platform combines Retrieval-Augmented Generation (RAG), Large Language Models (LLMs), and a multi-agent workflow to assist healthcare professionals with patient triage, evidence-based clinical guidance, and automated clinical documentation. The system is intended to support—not replace—clinical decision making. It generates recommendations based on patient information and medical knowledge while leaving final decisions to qualified healthcare professionals. 2. Key Features Multi-agent AI workflow Clinical triage support Evidence-based knowledge retrieval (RAG) Automated clinical note generation Offline-capable local LLM inference REST API Modern React frontend Modular architecture PostgreSQL / SQLite support ChromaDB vector search 3. System Architecture reports/system_architecture_diagram docs/images/system_architecture_diagram 4. Technology Stack Component Technology Frontend React + TypeScript + Tailwind CSS Backend FastAPI AI Workflow LangGraph Language Model Phi-3 / Llama (Ollama) Vector Database ChromaDB Database PostgreSQL / SQLite ORM SQLAlchemy Documentation pdoc / Sphinx Testing Pytest Deployment Docker (planned) 5. Repository Structure CDS-MAS/ ├── ai/ │ ├── agents/ │ ├── models/ │ ├── services/ │ ├── workflows/ │ └── rag/ │ ├── backend/ │ ├── app/ │ │ ├── api/ │ │ ├── database/ │ │ ├── models/ │ │ └── main.py │ ├── frontend/ │ ├── tests/ │ ├ │ ├── docs/ │ └── README.md 6. Installation Clone git clone github.com Create environment python -m venv newvenv Activate Windows newvenv\Scripts\activate Install Dependencies: requirements.txt pip install -r requirements.txt 7. Running the Backend uvicorn backend.app.main:app --reload Swagger localhost 8. Running the Frontend cd fr …