Using AI with mwc chat

In this course, you have access to the mwc chat feature, allowing you to use Generative AI in your labs and projects. This feature is highly experimental, so we're not sure what the experience will be like. We would love it if you could share your feedback with Chris over Discord or email, and tell us what you would like to see in the future.

Getting Started

Open a terminal and update MWC

uv tool upgrade making-with-code-cli

Then check that the installed version is 5.7.1 or later

mwc version

Start `mwc chat`

First, navigate to a project or a lab directory. For example,

cd ~/Desktop/making_with_code/mwc1/unit3/lab_tic_tac_toe

Then run

mwc chat

The first time you run this, mwc will install Pi, an AI coding harness. The installation process will ask you a few questions, press "y" to confirm and proceed with the process each time.

What is a coding harness? A coding harness is a program that connects an AI model to your computer so the AI can do more than chat. It can read your files, suggest changes, and run commands, while you watch and approve what it does. Many harnesses are TUIs: text user interfaces. A TUI runs inside the terminal and it has no buttons or menus. Your main form of interaction is by typing commands.

Using mwc chat

You can type in instructions or questions once mwc chat is running, and you can also ask it to look up information on the lab website if you provide the url.

Resume or reopen a chat

To reopen a chat from a previous session, type /resume and select the session you want to resume.

Exit a chat