Exercise 03: Your First AI-Generated Project — The Agentic Crew Hands-On Guide

Overview

So far you've set up your tools and submitted your first pull request. Now you're going to do something genuinely impressive: take a real open-source project, describe what you want in plain English, and let an AI agent transform it — generating custom illustrations, narration audio, and animated backgrounds using AI models from Hugging Face.

No prior coding experience required. The agent writes the code. You describe the vision.

Make your bullet points visual. Each one will become an AI-generated image, so specific and concrete works better than abstract. 'A warehouse filled with rows of server racks' will generate a better illustration than 'technology infrastructure.'
— from the exercise

What you'll build

By the end of this exercise

✓ A forked copy of the Web Presenter project

✓ Ten custom AI-generated illustrations using FLUX.1-schnell

✓ Spoken narration audio generated with text-to-speech models

✓ A customised Three.js animated background matching your colour palette

✓ A complete presentation you can open in any browser

Key concepts

Terms you'll learn

Term What it means

Hugging Face A platform hosting thousands of AI models — like GitHub but for machine learning

API token A secret key that proves your identity when making API calls

Text-to-image AI models that generate illustrations from plain English descriptions

Text-to-speech AI models that convert written text into spoken audio

Environment variable A value stored in memory — keeps secrets out of your code files

What's covered

Sections in this exercise

01 What You'll Build

02 What You'll Need

03 Set Your Hugging Face Token

04 Fork and Clone the Project

05 Explore the Project

06 Choose Your Topic

07 Generate the Illustrations

08 Generate the Narration

09 Update the Presentation

10 Customise the Animated Background

11 Preview in Your Browser

12 Commit and Push

What just happened

You described a vision in plain English and an AI agent turned it into a working product — custom illustrations, spoken narration, and animated backgrounds. You didn't write the code. You directed the agent. That's the shift.