This course is asynchronous, meaning there are no required live events, and you can complete activities at your own pace throughout each week.

The material is organized into five modules covering various topics through videos, readings, and hands-on exercises, and discussion forums:

Introduction Module: Pre-course orientation

Get your tools installed, your terminal working, and your GitHub account ready before the course starts. This orientation includes a welcome video from the instructor explaining what to expect and how the course is structured, along with a written setup checklist. There are no graded assignments during this stage—just instructions to help you install the required tools and prepare your environment before Week 1 begins.

Module 1: From chat window to command line

This module covers the two foundational skills of the course: getting into the terminal and building a persistent project environment. You’ll learn why CLI tools give you access to your machine’s computing power and why the same AI model behaves differently in a chat window than in a CLI tool. You’ll also set up a context file for a project so the AI knows your beat, sources, and standards before every session. By the end of the module, you’ll have a working CLI setup, a CLAUDE.md file you wrote yourself, and experience processing real documents from your filesystem.

Module 2: Custom skills for Claude Code

Once you have a working project environment, the next step is encoding your editorial standards into reusable tools instead of typing them every session. In this module, you’ll learn how to create custom slash commands, set up hooks that automate quality checks, and manage your CLI sessions effectively. The goal is to shift from crafting prompts on the fly to building a toolkit that reflects how your newsroom works.

Module 3: CLI workflows for newsrooms

This module focuses on automation—using AI to build scripts and pipelines that handle recurring newsroom tasks. Instead of building automation by hand, you’ll describe the workflow in plain English and let the AI construct the process. You’ll also learn when automation makes sense for a task, how to inspect and understand what the AI built, and why testing small before running large jobs is important when working with batch processing.

Module 4: Agents, tools, and data access

The final module covers how AI systems take multi-step actions, connect to data sources, and where these connections break down. You’ll learn the difference between chatbots and agents, explore interactive and non-interactive modes, and connect AI tools to local files and external data using the Model Context Protocol (MCP). You’ll also examine common points where data connections fail—such as permissions, authentication, and schema mismatches—and how to debug them.

Register now and gain immediate access to the introduction module materials. If you have any questions, please contact us at journalismcourses@austin.utexas.edu.