Cortex

A Secure Document-based Q&A System

🚀 Project Overview

Cortex is a secure Retrieval-Augmented Generation (RAG) system built using LangChain, OpenSearch and HuggingFace embeddings. It enables multi-document semantic search, private data handling and intelligent question suggestions. Prioritized reliability, scalability and usability. For me, Cortex isn’t just a side project – it’s about showing how AI can be built responsibly, with security and clarity at its core.

🔑 Key Features

🔒 Data stays private – only your uploaded files are used.

📂 You choose which documents to include in retrieval, giving full control over context.

📚 Cortex handles multiple documents for richer context and better answers.

📝 Accurate answers, grounded solely in the source document.

💡 Smart follow-up question suggestions to guide deeper exploration.

⚡ Tech Stack

• LangChain

• OpenSearch

• HuggingFace Embeddings

• Python

🎥 Demo

CortexV1.0.mp4

📚 Next Steps

Version 2 planned -

• Smarter reference linking
• Improved retrieval experience
• Data mapping enhancements