Welcome to CodeTeach.ai
First time here? In 10 minutes you can have your first AI-generated assignment deployed to your GitHub Classroom. This article shows you what to expect and where to go next.
Written By Alan Gandy
Last updated About 1 month ago
If you've been handed CodeTeach.ai and told "build your assignments here," this is the right place to start. In about 10 minutes you can take a one-paragraph description of what you want students to build and walk away with a complete assignment — starter code, solution, autograder tests, student instructions, and a deployed GitHub Classroom repo.
This article is the orientation. It tells you the shape of the workflow, what you'll need on hand, and which article to read next.
What you're about to do
Three steps, repeated for every assignment you make:
- Seed — Click New Assignment in the sidebar. Pick what kind of input you have (existing code, a description, or both). The AI generates four artifacts: starter code, solution code, autograder tests, and student-facing instructions.
- Review & Edit — Read what the AI produced. Edit anything that needs polish in an in-browser editor. Approve the instructions. Run the autograder check, which simulates what GitHub Classroom will run on student submissions.
- Deploy — Push to a GitHub repository in your organization. CodeTeach walks you through wiring it into GitHub Classroom so students can join.
That's it. Every assignment passes through those three stages. The progress bar across the top of the wizard always shows you which stage you're in.

When you sign in, that's what you see. The sidebar holds your six destinations (Dashboard, New Assignment, Credits, Refer & Earn, Settings, Help). The main panel lists every assignment you've created and what state each is in. Every row is clickable — opening one drops you back into whichever stage that assignment is in.
What you need on hand before you start
Three things, in order of how often instructors get stuck on them:
A GitHub organization — not a personal account. GitHub Classroom can't host template assignments from personal accounts (the repos are either invisible to Classroom or expose your solution branch publicly). If you don't already own an org, the Setting up GitHub for CodeTeach article walks through creating one.
The CodeTeach GitHub App installed on that organization. CodeTeach uses the App to create the assignment repo and push starter + solution branches. You'll do this once per org; we'll prompt you on the Deploy step if it isn't installed yet.
GitHub Classroom activated on that organization. Free for verified educators at classroom.github.com. If you've never used Classroom before, do this next — verification can take a day, and the rest of CodeTeach won't help you ship to students until you have a Classroom.
Where to go next
- First time? Never used GitHub Classroom? → Read Setting up GitHub for CodeTeach next. It walks through the org + App + Classroom prerequisites end-to-end.
- Got the prerequisites, ready to make your first assignment? → Read Picking a generation mode. It explains the four ways to seed an assignment so you can pick the right one for what you have on hand (code, prose, or both).
- Already in the middle of an assignment that's hit a snag? → Open it from your dashboard. The wizard reopens at whichever stage it's in. If validation failed or deploy failed, the on-page banner tells you what to do.
- Want to see what students will experience? → After you deploy your first assignment, GitHub Classroom gives you an invite URL. Test-drive it yourself with a personal GitHub account before sharing with students.
A note on the difficulty level
When you create an assignment, you'll pick Beginner, Intermediate, or Advanced. This is more consequential than it looks — it doesn't just adjust code complexity, it directly controls how much working code appears in the student-facing Instructions.md. Beginner gives students complete code samples for every TODO; Advanced gives them only requirements with no code at all. Pick wrong and the assignment is the wrong difficulty regardless of how clever the code is. Choosing a difficulty level explains the trade-offs.