RAGe!

RAGe is an AI-powered research analysis platform that helps scholars and innovators instantly discover limitations and future directions in academic research.

Comment

Inspiration

Keeping up with the latest research is overwhelming, thousands of papers are published every week across multiple fields. We wanted to create a tool that helps researchers quickly identify gaps, limitations, and future directions without manually combing through endless articles.

What it does

RAGe is a multi-layer AI research platform that fetches papers from arXiv, Semantic Scholar, PubMed, and Crossref, performs a two-stage hybrid search (keyword filtering + semantic re-ranking), and uses Claude AI to extract research insights. It uses these methods to quickly parse relevant sections of research papers and return data which it then uses to make interactive visualizations and a function-calling chatbot to explore papers in real time, giving users a full picture of trends, gaps, and opportunities.

How we built it

We combined several technologies into a unified platform: Elasticsearch for fast keyword-based retrieval, ChromaDB for semantic embedding and re-ranking, Claude AI for high level research gap analysis and chatbot interactions, Streamlit for an interactive, visually appealing dashboard, Multi-source fetching ensures at least 5 recent papers per query, with local caching for speed.

Challenges we ran into

Multi-source integration: APIs had different formats and rate limits, requiring careful orchestration. Hybrid search optimization: Combining keyword and semantic search to maximize precision while keeping performance fast. We had to fine tune this system to show and parse through modern/highly relevant data. Real-time AI analysis: Balancing speed with accurate extraction of research limitations and future directions.

Accomplishments that we're proud of

A fully functioning two-stage hybrid retrieval system with multi-source fallback. Seamless integration of Claude AI for actionable research insights. An interactive dashboard with live trend visualizations and a Q&A chatbot. Ensuring recent paper prioritization and a minimum paper guarantee for every query. Creating a ElasticDB that has parsed through over 1k+ research papers.

What we learned

Building a research assistant is both a data engineering and AI challenge, handling structured and unstructured data at scale requires careful design. Multi-stage retrieval significantly improves precision and relevance over standalone searches. Users value actionable insights, not just summaries, highlighting the importance of research gap analysis.

What's next for RAGe!

Expand support for data collection through more academic databases than just our main 4 to cover broader disciplines. Add automated literature review reports summarizing key gaps, trends, and references. Optimize the platform for collaborative research workflows, enabling teams to explore and annotate papers together.

Built With

Share this project:

Updates