paper-qa

Why we included this project

Anyone who regularly reads through PDFs and preprints will find this worth a look. PaperQA2 turns a pile of papers, text files, Office documents, and even source code into a searchable knowledge base that answers questions with line-level citations back to the originals. Its retrieval pipeline is tuned for scientific literature, so it holds up on the kind of multi-step questions where you want a defensible, cited answer rather than a confident guess. Because it works both as a command-line tool and a Python library, it drops cleanly into existing analysis scripts or ad hoc research workflows. If you need traceable answers from your own document collection, not just a generic chatbot, this is a practical fit.