Researchers introduced Paper2Agent, a framework that turns research papers into AI agents. Each agent acts as a virtual corresponding author for its paper. It exposes manuscripts, code, data and workflows as active knowledge. The system builds a model context protocol server from the paper and codebase. Tests refine and increase the robustness of the server. These servers connect to chat agents like Claude Code. Users can run complex scientific queries in natural language. Case studies show agents reproducing original results and answering new queries. Multiple agents can collaborate to generate novel insights. The approach aims to create a collaborative ecosystem of AI co-scientists.