Feynman Reviews — Discover what people think of this product.

Feynman is an open source AI research agent that runs entirely on your machine, helping you move from vague questions to fully cited scientific outputs. It reads papers, searches the web, audits code, and drafts publication-ready documents while keeping every claim grounded in primary sources.

With Feynman, you can orchestrate multi-agent workflows for deep scientific investigation, from mechanistic interpretability dives to RLHF alternative surveys. Use natural language or slash commands to trigger:

• /deepresearch for multi-agent investigations across papers, web, and repositories
• /lit for structured literature reviews with consensus and open questions
• /audit for paper-to-code reproducibility checks in isolated Docker environments
• /replicate for experiment plans and execution on local or burst GPU compute

Feynman integrates with alphaXiv, Pi skills, and managed GPU providers like Modal and RunPod, giving you a reproducible research pipeline that stays local-first and source-cited. Whether you’re validating scaling laws, planning replication studies, or drafting peer-review-ready manuscripts, Feynman assembles specialized agents—Researcher, Reviewer, Writer, Verifier—to deliver transparent, verifiable outputs.

Designed for scientists, engineers, and research-heavy teams, Feynman transforms unstructured questions into rigorous research artifacts, complete with inline citations, agreement matrices, and revision plans. Install it via a simple script, keep your workflows auditable, and let autonomous research loops watch for new papers, code, and product updates.