Stanford's Paper2Agent Turns Research Papers Into Reproducible AI Agents
Computational research papers typically ship with code that readers must clone, install, configure, and debug. That overhead keeps useful methods locked inside PDFs. A Stanford team led by Jiacheng Miao and James Zou proposes a fix. Paper2Agent was published in Nature on September 16, 2026. It converts a paper and its codebase into a Model Context Protocol (MCP) server. Any MCP-compatible agent, such as Claude Code, can then run the paper's methods through natural language. The authors describe the result as a virtual corresponding author.
In 2026, as agentic tooling matures, Paper2Agent tackles a persistent bottleneck: research reproducibility. By packaging methods as MCP servers, it enables AI agents to execute paper-specific workflows directly, reducing the manual effort required to reuse computational findings.
Is It Deployable?
Yes. The code is MIT-licensed and installs as a skill for Claude Code or Codex. Prebuilt AlphaGenome, Scanpy, and TISSUE servers run on Hugging Face Spaces. A hosted version is also available at paper2agent.ai.
How the Pipeline Works
Paper2Agent runs on Claude Code's agent SDK. A central orchestrator dispatches specialized sub-agents through six steps:
- Locate and download the codebase.
- An environment manager builds an isolated virtual environment.
- A tutorial scanner indexes usable tutorials.
- A tutorial executor runs them end to end and records results.
- (Additional steps continue as described in the original pipeline.)
By automating environment setup and tutorial execution, Paper2Agent lowers the barrier to reproducing and extending computational research.
via MarkTechPost