MedIntel-RAG is an AI-powered Retrieval-Augmented Generation (RAG) system built to assist healthcare professionals in accessing reliable, up-to-date medical knowledge from The Merck Manuals. It helps streamline clinical decision-making, reduce information overload, and improve patient care outcomes.
- Diagnostic Assistance: Get key symptoms, signs, and treatments for various conditions.
- Drug Information: Look up trade names, usage, and dosages for medications.
- Treatment Plans: Retrieve evidence-based first-line and alternative treatment options.
- Specialty Knowledge: Access structured diagnostic steps for specialty areas like endocrinology.
- Critical Care Protocols: Quickly access emergency care protocols like those for sepsis.
- Retrieval-Augmented Generation (RAG)
- Vector Store (e.g., FAISS) for semantic search
- Large Language Model (e.g., GPT-4 or similar)
- PDF Parsing & Chunking (Merck Manual)
- Frontend UI (Streamlit or FastAPI)
- The Merck Manuals (PDF, over 4,000 pages, 23 sections)
- Used as the trusted, centralized medical knowledge base.
- Chunked and indexed for semantic retrieval.
Note: Ensure usage complies with copyright and licensing terms.
git clone https://github.com/your-username/MedIntel-RAG.git
cd MedIntel-RAG
pip install -r requirements.txt