Overview

Feynman is your open-source AI research agent, designed to streamline and accelerate your scientific and technical exploration. It empowers you to interact with vast amounts of information, synthesize findings, and even generate research artifacts, all through an intuitive command-line interface and a comprehensive workbench.

Key Features

• Effortless Installation: Get started quickly with simple one-line installers for macOS, Linux, and Windows, or opt for an npm alternative.
• Powerful Research Commands: From quick summaries and deep multi-agent investigations to literature reviews, paper analysis, and replication planning, Feynman offers a rich set of tools.
• Intelligent Agentic Workflows: Leverage specialized agents like the Researcher, Reviewer, Writer, and Verifier to automate complex research tasks.
• Extensive Skill & Tool Integration: Seamlessly connect with AlphaXiv for paper search, Hugging Face Hub for dataset insights, and a vast array of scientific databases and tools via Feynman Bio Tools.
• Comprehensive Science Workbench: Access a local, standalone workbench for project management, in-app chat, notebook integration, artifact previews, provenance tracking, and more.
• Local Model Support: Easily integrate with popular local LLM platforms like LM Studio, LiteLLM, Ollama, and vLLM for a personalized AI experience.
• Reproducible Research: Plan and execute replication checks, audit claims against public codebases, and find implementable ML training recipes.
• Grounded Outputs: All research outputs are source-grounded, with claims directly linked to papers, documentation, or repositories.

Typical Use Cases

• Academic Research: Quickly gather and synthesize information on a topic, identify seminal papers, and understand the state-of-the-art.
• Software Development: Research new libraries, understand complex APIs, and find best practices for implementation.
• Technical Due Diligence: Analyze papers and codebases to assess novelty, reproducibility, and potential claims.
• Machine Learning Engineering: Discover and evaluate ML training recipes, reproduce experimental results, and identify optimal models.
• Scientific Exploration: Dive deep into specific research areas, explore literature trends, and generate hypotheses.
• Knowledge Management: Organize and revisit past research sessions, retrieve information efficiently, and maintain a clear provenance trail.