feynman
Feynman is an open-source AI research agent designed to help developers and researchers investigate scientific literature, audit papers, and synthesize research findings. Built on the Pi agent runtime and integrated with alphaXiv for paper search, it provides specialized workflows for ranking papers, performing literature reviews, replicating experiments, and generating ML training recipes. The agent can be deployed as a standalone terminal application with a science workbench interface, or its research skills can be installed separately into Codex, Claude, or OpenCode environments.
Key Features
Use Cases
- 01Ranking and prioritizing which research papers to read first on a specific topic using citation and methodology evidence
- 02Conducting comprehensive literature reviews with consensus findings, disagreements, and open questions across a research area
- 03Auditing machine learning papers by comparing published claims against the actual codebase implementation
- 04Finding and ranking implementable ML training recipes from papers, datasets, documentation, and code repositories
- 05Setting up research watches to establish baselines on topics with optional scheduled follow-ups for new developments
- 06Extracting and reusing specialized research skills (paper search, biomedical database queries, chemistry tools) in existing Codex or Claude agent workflows
Related Agents
View morehermes-agent
The agent that grows with you
agency-agents
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
openinterpreter
A coding agent for open models like Kimi K3
cline
Autonomous coding agent as an SDK, IDE extension, or CLI assistant.
feynman — FAQ
What is Feynman?+
Feynman is an open-source AI research agent that helps developers and scientists search academic papers, perform literature reviews, audit research claims, and synthesize findings. It can run as a standalone terminal application or provide skills for other agent frameworks like Codex and Claude.
How do I install Feynman?+
On macOS or Linux, run 'curl -fsSL https://feynman.is/install | bash' to install the full terminal application. On Windows, use PowerShell: 'irm https://feynman.is/install.ps1 | iex'. For skills-only installation into Codex or other agents, use the install-skills script with appropriate scope flags (--codex, --repo, or --opencode).
Which AI clients work with Feynman?+
Feynman runs as a standalone terminal agent with its own Pi runtime. Its research skills can be extracted and installed into Codex (via ~/.codex/skills/feynman), Claude/agent repositories (via .agents/skills/feynman), or OpenCode projects (via .opencode/skills/feynman).
Do I need API keys to use Feynman?+
Feynman supports local models (LM Studio, Ollama, vLLM, LiteLLM) without API keys via 'feynman setup'. For cloud providers, you'll need to authenticate using 'feynman model login <provider>' or set environment variables like OPENROUTER_API_KEY. Hugging Face Hub tools optionally use HF_TOKEN for enhanced access.
Is Feynman free to use?+
Yes, Feynman is open-source software released under the MIT License. However, you may incur costs if you choose to use paid language model providers (OpenRouter, cloud APIs) instead of local models.
What are the main workflows Feynman supports?+
Feynman provides workflows for ranking papers (feynman rank), conducting deep multi-agent research (feynman deepresearch), performing literature reviews (feynman lit), auditing paper claims (feynman audit), planning replication studies (feynman replicate), and finding ML training recipes (feynman recipe). Each workflow produces source-grounded outputs with direct citations.
How do I install feynman?+
Open the source repository on GitHub and follow its README. feynman is a agent — MCP Agents Market links you directly to the official repo.
Is feynman free?+
feynman is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.