Future-House/paper-qa

High accuracy RAG for answering questions from scientific documents with citations

What it solves

PaperQA2 is designed to perform high-accuracy, agentic retrieval-augmented generation (RAG) on scientific literature. It solves the challenge of extracting grounded, cited answers from large collections of PDFs, text files, Microsoft Office documents, and source code by automating the search, evidence gathering, and synthesis process.

How it works

The system follows a three-phase algorithm: first, it searches for candidate papers using LLM-generated queries; second, it chunks, embeds, and ranks document passages to gather relevant evidence; finally, it uses an LLM to synthesize a grounded answer with in-text citations. It supports agentic workflows where an LLM can iteratively refine its search and evidence gathering, and it includes automatic metadata retrieval from sources like Semantic Scholar and Crossref.

Who it’s for

It is for researchers, scientists, and developers who need to perform complex question answering, summarization, or contradiction detection across large repositories of technical documents and scientific papers.

Highlights

  • Grounded responses that include in-text citations to source documents.
  • Agentic RAG capabilities allowing for iterative query refinement.
  • Support for diverse file types including PDFs, Office documents, code, and multimodal content like tables and figures.
  • Built-in CLI for rapid indexing, searching, and querying of local document directories.
  • Highly configurable with bundled settings for different performance needs (e.g., "fast" vs. "high_quality").
  • Compatibility with a wide range of LLM providers via LiteLLM.

Related

  • Project
    BenyRonald77/uajy-academic-rag-chatbot

    A Retrieval-Augmented Generation (RAG) chatbot designed for UAJY academic documents that uses hybrid retrieval and LLM reranking to provide grounded, cited answers.

  • Project
    AutoLab-SAI-SJTU/Paper2Rebuttal

    An AI-powered multi-agent system that helps researchers analyze reviewer comments, search for supporting literature, and draft formal academic paper rebuttals.

  • Project
    SaiAkhil066/CORTEX-AI-SUPER-RAG

    Cortex AI SUPER RAG is a local, open‑source RAG app that lets you upload PDFs/Word docs, ask natural‑language questions, and receive answers with precise citations. It combines BM25 keyword search, FAISS vector search, optional reranking, entity‑graph expansion, and several LLM‑based tricks—all runnable via Ollama without cloud APIs. The repo includes a Streamlit UI, a full evaluation harness (e.g., FinanceBench benchmark), and Docker support.

  • Project
    NirDiamant/Controllable-RAG-Agent

    A RAG agent that uses a deterministic decision graph to plan, execute sub-tasks, retrieve and distill evidence, verify grounding against the source documents, and iteratively re-plan — built to answer multi-hop questions a simple semantic search can't handle.

  • Project
    Paper2Poster/Paper2Poster

    Paper2Poster is a multimodal automation system that converts scientific papers (PDF) into editable presentation posters (PPTX) using a multi-agent pipeline, and provides an evaluation benchmark (PaperQuiz, VLM-as-Judge, statistical metrics) to assess poster quality.