Getting Started

What you need before Month 1 begins, the one-time setup, and the AI tools the tutor supports.

You’ll need

  • A Mac (Apple Silicon or Intel) running macOS.
  • A free GitHub account.
  • About 8–12 hours per week if you’re following the 12-month pace. (Or 15–20 hrs/week for the 8-month compressed pace, or full-time for 5–6 months.)
  • One AI coding tool installed — see the supported tools below. You can use any of them, including a $0 local-model setup.

You do not need to know how to code. The first five months teach you that.

Clone the course 

git clone https://github.com/lernen-edu/agentwright.git
cd agentwright

That’s it — no setup script to run. Open the directory in your AI coding tool of choice (see below) and the tutor activates automatically.

Install the macOS baseline

Every month uses the same free tooling. Install Homebrew first, then the rest in one go:

# Homebrew (if you don't have it):
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

# Core stack
brew install git gh uv httpie tmux
brew install --cask visual-studio-code

# Free LLM (used in every AI month)
brew install ollama

Don’t worry about installing everything up front — each month tells you what new tool it needs.

Supported AI tools

The Agentwright tutor activates automatically when you open this directory in any of these:

  • Claude Code — Anthropic’s CLI. Free tier exists; the $20/mo Claude.ai plan also includes Claude Code. Reads CLAUDE.md.
  • OpenAI Codex CLI — OpenAI’s CLI for code. Reads AGENTS.md.
  • Gemini CLI — Google’s CLI. Free tier available. Reads GEMINI.md.
  • Opencode — open-source coding agent. Reads AGENTS.md.
  • Pi Coding Agent — reads AGENTS.md.

You can also bring your own local model setup via Ollama and a wrapper tool. The tutor instructions live in plain markdown files at the root of this repo, so any tool that reads agent-context conventions will pick them up.

What the tutor will and won’t do

Read How the Tutor Works before you start. The short version: it’s Socratic by default. It asks you questions, points you at course sections, and gives hints with a gap. It will not write your lab code or run your commands. That is the point.

If you ever need a quick factual answer, say “answer me directly” — you’ll get one direct answer, then the tutor returns to Socratic mode.

Track your progress

The tutor reads .tutor/progress.md at the start of every session to calibrate how much scaffolding to give you. Update it as you move through the course. The lab work is yours; the progress log is also yours.

When you’re ready

Open the curriculum index and start at Month 1, Lab 1. Work top to bottom. Hit each month’s Month-End Assessment before moving on.

Good luck.