paper-qa and msmarco-genqa

These are complements: paper-qa provides a complete production-ready RAG system for scientific documents, while msmarco-genqa demonstrates core RAG components (BM25 retrieval, FAISS indexing, reranking) that could be integrated into or compared against paper-qa's architecture.

paper-qa
77
Verified
msmarco-genqa
24
Experimental
Maintenance 20/25
Adoption 12/25
Maturity 25/25
Community 20/25
Maintenance 13/25
Adoption 2/25
Maturity 9/25
Community 0/25
Stars: 8,264
Forks: 838
Downloads:
Commits (30d): 7
Language: Python
License: Apache-2.0
Stars: 2
Forks:
Downloads:
Commits (30d): 0
Language: Jupyter Notebook
License: MIT
No risk flags
No Package No Dependents

About paper-qa

Future-House/paper-qa

High accuracy RAG for answering questions from scientific documents with citations

Implements agentic RAG with iterative query refinement and LLM-based re-ranking, automatically enriches documents with metadata (citations, journal quality) from Semantic Scholar and Crossref, and supports multiple document formats (PDFs, text, code, Office files) with full-text search via tantivy. Integrates with any LiteLLM-supported model provider and offers local embedding alternatives, enabling deployment without proprietary APIs.

About msmarco-genqa

GioiaZheng/msmarco-genqa

RAG-based search question answering system on MS MARCO with BM25 retrieval, FAISS indexing, and transformer reranking.

Scores updated daily from GitHub, PyPI, and npm data. How scores work