Rajni Research Assistant: An Intelligent Academic Paper Analysis Assistant Based on RAG

Introducing the Rajni Research Assistant open-source project, an intelligent research assistant based on Streamlit, LangChain, ChromaDB, and large language models. It aims to address the challenges of literature processing in academic research, enabling paper analysis, knowledge extraction, and intelligent Q&A through RAG technology. The project is open-sourced on GitHub and maintained by rajnitiwari28.

In the field of academic research, researchers need to read a large number of papers to keep up with cutting-edge developments, but the massive volume of literature makes it challenging to efficiently extract information and establish knowledge connections. In recent years, RAG technology has emerged, providing new ideas to solve this problem. Rajni Research Assistant is exactly an intelligent assistant for academic scenarios based on RAG technology.

The project's technical architecture includes: Streamlit as the front-end framework to provide an intuitive interface; LangChain to coordinate component interactions, simplifying prompt management and chain calls; ChromaDB as a vector database to store document embeddings and support semantic retrieval; and support for large language models like Gemini and Groq for summarization and Q&A. The RAG principle is: convert user queries into vectors → retrieve similar document fragments → construct enhanced prompts → model generates answers, improving factuality and interpretability.

Core functions include: 1. Paper upload and vectorization: support PDF upload, automatically extract text chunks, embed them, and store in ChromaDB; 2. Intelligent Q&A and summarization: generate accurate answers based on RAG-retrieved relevant context to reduce model hallucinations, and automatically generate paper summaries; 3. Cross-paper knowledge association: support multi-document Q&A and comparative analysis to help build knowledge graphs.

Application scenarios are wide-ranging: graduate students/PhD candidates can use it to assist with literature reading and review; research teams can build shared knowledge bases to avoid information silos; journal editors/reviewers can quickly understand the relationship between submissions and existing literature; enterprise R&D can build internal technical document knowledge bases to support research and innovation.

Limitations: limited PDF format support (weak OCR for scanned versions), difficulty in recognizing complex charts and formulas, chunking strategy affecting retrieval quality, and multi-language support needing optimization. Improvement directions: introduce advanced document parsing technology, support multi-modal content understanding, optimize retrieval reordering, and enhance citation tracing functions.

Rajni Research Assistant is an excellent open-source example of RAG technology applied in academic scenarios. It integrates multiple technologies to provide researchers with practical tools, and also serves as a learnable and extensible RAG application template. With the advancement of AI technology, such tools will play a greater role in academic research and knowledge management.