Full‑stack RAG assistant for market‑access evidence queries
The project provides a Docker‑based full‑stack application that lets users upload PDFs, TXT or DOCX evidence and ask citation‑backed questions using Retrieval‑Augmented Generation. It includes a FastAPI backend, a React/Vite frontend, PostgreSQL storage and optional Gemini integration for real embeddings and LLM responses. Designed for market‑access analysts, it offers mock mode for offline testing and a live mode for production‑grade answers. Compared to ad‑hoc scripts, it delivers a ready‑to‑run, end‑to‑end solution with authentication and source tracing.
View on GitHub →martinssplendour/Rag-System