Smart-Paper-Analyst

Smart Paper Analyst is an AI-powered academic research assistant that uses Retrieval-Augmented Generation (RAG) to help users understand and analyze research papers. It allows users to upload research papers in PDF format and interact with their content through semantic search, AI-generated answers, summaries, paper comparison, and research-gap analysis. The system combines document processing, vector-based retrieval, and Large Language Model (LLM) generation to provide context-aware responses based on the uploaded papers. The application follows a RAG-based pipeline: PDF research papers are processed through text extraction, text chunking, embeddings generation, FAISS vector store, semantic retrieval, relevant context, and LLM generation to produce AI-powered responses. The project uses the Sentence Transformers all-MiniLM-L6-v2 model to generate normalized vector embeddings for document chunks and queries, and the generated embeddings are indexed using FAISS IndexFlatIP, enabling similarity-based retrieval. Retrieved document context is passed to a configured LLM to generate the final response, and the project supports multiple LLM providers through a unified client: OpenRouter, Mistral API, and Hugging Face Inference API.