Paper2Agent: Reimagining Research Papers As Interactive and Reliable AI Agents
Jiacheng Miao, Joe R. Davis, Jonathan K. Pritchard, James Zou
2025-09-09
Summary
This paper introduces a new system called Paper2Agent that automatically turns research papers into interactive AI assistants. Instead of just reading a paper, you can essentially *talk* to it and get it to perform tasks described in the research.
What's the problem?
Currently, using research is hard. Scientists spend a lot of time trying to understand someone else's code, data, and methods from a paper to apply it to their own work. This slows down progress because it's difficult to build upon existing research and reuse what others have already done. Papers themselves are static and don't actively help you use the information inside.
What's the solution?
Paper2Agent solves this by analyzing a research paper and its associated code. It uses a bunch of different AI 'agents' working together to create a special system, called a Model Context Protocol (MCP), that understands the paper's content. Then, it tests and improves this system to make sure it's reliable. Finally, it connects this system to a chat agent, like Claude Code, so you can ask questions in plain English and the AI will use the paper's tools and methods to answer them. They tested it with papers on genetics and analyzing cells, and it worked!
Why it matters?
This is a big deal because it changes how scientific knowledge is shared and used. Instead of papers being something you just read, they become dynamic AI tools you can interact with. This could lead to faster discoveries and a more collaborative environment where AI helps scientists build on each other's work, essentially creating 'AI co-scientists'.
Abstract
We introduce Paper2Agent, an automated framework that converts research papers into AI agents. Paper2Agent transforms research output from passive artifacts into active systems that can accelerate downstream use, adoption, and discovery. Conventional research papers require readers to invest substantial effort to understand and adapt a paper's code, data, and methods to their own work, creating barriers to dissemination and reuse. Paper2Agent addresses this challenge by automatically converting a paper into an AI agent that acts as a knowledgeable research assistant. It systematically analyzes the paper and the associated codebase using multiple agents to construct a Model Context Protocol (MCP) server, then iteratively generates and runs tests to refine and robustify the resulting MCP. These paper MCPs can then be flexibly connected to a chat agent (e.g. Claude Code) to carry out complex scientific queries through natural language while invoking tools and workflows from the original paper. We demonstrate Paper2Agent's effectiveness in creating reliable and capable paper agents through in-depth case studies. Paper2Agent created an agent that leverages AlphaGenome to interpret genomic variants and agents based on ScanPy and TISSUE to carry out single-cell and spatial transcriptomics analyses. We validate that these paper agents can reproduce the original paper's results and can correctly carry out novel user queries. By turning static papers into dynamic, interactive AI agents, Paper2Agent introduces a new paradigm for knowledge dissemination and a foundation for the collaborative ecosystem of AI co-scientists.