MarketMuse

is a Retrieval-Augmented Generation (RAG) system designed to assist content strategists in analyzing and extracting insights from their document collections. Built with robust security protocols and ethical AI guidelines, this assistant provides document-based question answering while preventing prompt extraction attacks and maintaining strict privacy standards.

The key features are :

• Secure Document Analysis: Processes uploaded documents (PDF, TXT, DOCX, PPTX, HTML, MD) with local embedding generation.
• Strict Contextual Integrity: Answers only based on uploaded document content, preventing hallucination
• Ethical AI Framework: Implements comprehensive safety guidelines and content boundaries

Technical Stack

• Streamlit UI
• LangChain RAG
• ChromaDB
• HuggingFace Embeddings
• Groq API

User Interface