Overview
ResearchRAG is an end-to-end Retrieval-Augmented Generation (RAG) system for answering questions over scientific papers. It retrieves semantically relevant paper passages using BGE embeddings and FAISS, then generates grounded answers with inline citations. The system is built as a portfolio project to demonstrate a practical RAG workflow covering document ingestion, text processing, vector search, source attribution, and retrieval evaluation. The architecture includes downloading papers from arXiv, parsing and cleaning PDF text, splitting documents into token-based chunks, embedding with BAAI/bge-small-en-v1.5, indexing with FAISS, retrieving candidate chunks, deduplicating at the paper level, building grounded prompts, generating answers with Gemini, and validating citations. It includes a small labeled retrieval smoke-test set evaluated with Hit@5 and Mean Reciprocal Rank (MRR).
Pricing
Free
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Researchers use ResearchRAG to ask questions over scientific papers and receive answers with inline citations.
- Developers use ResearchRAG to download scientific papers from arXiv while preserving metadata.
- Data scientists use ResearchRAG to parse and clean PDF text, remove reference sections, and split documents into overlapping token chunks.
- Engineers use ResearchRAG to build a FAISS index with BGE embeddings for semantic retrieval.
- Analysts use ResearchRAG to evaluate retrieval performance with Hit@5 and Mean Reciprocal Rank (MRR).
- Teams use ResearchRAG to generate grounded answers that avoid unsupported claims and include inline citations.
- Users use ResearchRAG to validate generated citation identifiers against retrieved contexts before displaying sources.
Features
Scientific paper question answering
BGE based text embeddings
FAISS based document retrieval
Gemini generated answers
Answer citation validation
Retrieval quality evaluation
Reviews
No reviews yet. Be the first to share your experience!