AutoResearch Guide
Unofficial guide
Home / AutoResearch GitHub Guide
Support Page

AutoResearch GitHub Guide: What It Is and Where to Start

If you searched for autoresearch github or karpathy auto research github, the repository most people mean is karpathy/autoresearch. This page explains what the repo is, why it matters, and which files to look at first.

What Is AutoResearch?

AutoResearch is an open-source experiment in autonomous ML research. It gives an AI coding agent a compact training loop and asks it to improve the setup by editing code, running short experiments, and checking whether the result got better.

The repo is small on purpose. That makes it easier to inspect than a large research stack and easier for agents to work with.

Where Is the Official GitHub Repo?

The official source is github.com/karpathy/autoresearch. This site is not the official home of the project.

Which Files Matter Most?

  • prepare.py for data prep, tokenizer setup, and utilities
  • train.py for the main training logic the agent is expected to change
  • program.md for the human-written instructions that guide the agent

FAQ

Who created AutoResearch?

Andrej Karpathy created the repository karpathy/autoresearch.

Is the GitHub repo the official source?

Yes. The GitHub repository is the official code source. This page is only an explanatory guide.

Where should a beginner start?

Start with the README, then review program.md and train.py so you understand the research loop.