Learn from the source
Everything below is published by the people who build these models and tools. Free, first-party, and worth more than any summary of it — including this one.
Official academies
Structured, free courses straight from the model makers. Start here, not with a YouTube playlist.
Anthropic Academy
Courses on Claude, prompting, the API, and Claude Code, written by the team that built them.
AnthropicOpenAI Academy
Free courses and live sessions from beginner ChatGPT use to building with the API.
OpenAIGrow with Google — AI
Practical AI training aimed at workers and small businesses, including AI Essentials.
Googlen8n Courses
The official beginner and advanced automation courses. Do these before Day 4.
n8nMeta Blueprint
Meta's own advertising curriculum, including campaign structure and measurement.
MetaPrimary documentation
When a tutorial and the docs disagree, the docs are right. Bookmark these on Day 1 and keep returning.
Prompt engineering overview
The techniques ranked by impact — be clear and direct, use examples, let the model think.
AnthropicBuilding effective agents
The essay that separates workflows from agents. Read it before Day 3, twice.
AnthropicPrompt engineering interactive tutorial
Nine chapters of hands-on prompting exercises with answers. The single best Day 1 homework.
AnthropicAnthropic Cookbook
Runnable notebooks: tool use, RAG, sub-agents, vision, evaluation.
AnthropicOpenAI Cookbook
Worked examples for the API, structured outputs, and evaluation.
OpenAIGemini prompting guide
Google's plain-language prompt framework: persona, task, context, format.
Googlen8n documentation
Node reference, expressions, error handling. Day 4 lives in here.
n8nEbooks and whitepapers from the model makers
Long-form, free, and written by the companies whose models you are using. Download them; they are worth more than most paid courses.
A Practical Guide to Building Agents
When an agent is the right answer, how to design one, and how to guardrail it.
OpenAIAI in the Enterprise
Seven lessons from real deployments. Useful framing for the Aimers corporate-training pitch.
OpenAIIdentifying and Scaling AI Use Cases
How to find the work in your own business that AI should touch first.
OpenAIPrompt Engineering (whitepaper)
Sixty-odd pages on prompt structure, configuration, and reasoning techniques.
GoogleAgents (whitepaper)
Google's take on agent architecture: models, tools, and the orchestration layer.
GoogleClaude Code best practices
How to actually work with an agentic coding tool. Relevant from Day 2 onward.
AnthropicBuilding with AI
Everything Day 2 asks you to install, plus the page to go back to when a step refuses to work.
Claude Code documentation
Install, configure, and work with Claude Code. Start at the overview, then the quickstart.
AnthropicClaude Code skills
What a skill is, where the folder lives, and how to write one of your own.
AnthropicConnecting an MCP server
How to attach an outside system to Claude Code, and how to check it actually connected.
AnthropicModel Context Protocol
The open standard behind MCP, and the directory of servers other people have published.
AnthropicLovable documentation
Prompting, editing, connecting GitHub, and publishing — the four things you need today.
LovableGitHub: Hello World
Repository, branch, commit, pull request — explained once, properly, in about fifteen minutes.
GitHubDeploy a static site to Cloudflare Pages
The exact path this class uses: connect the repository, leave the build command empty, publish.
CloudflareUI UX Pro Max
Design skill for Claude Code. The install commands are in the readme, and in section 13 today.
nextlevelbuilderAcademic Research Skills
Research skill with a deep-research mode. Used today to check claims before they reach a page.
Imbad0202Motion for React (Framer Motion)
The animation library behind the motion work today. Read the reduced-motion section too.
MotionHardik Pandya
stop-slop is not on a marketplace — you create the folder yourself. Section 13 has the steps.
Author of stop-slopGlossary
The terms that come up across both paths, in both languages.
LLM
A model trained to predict the next piece of text. Everything else it appears to do follows from that.
Prompt
The full instruction you give a model: role, context, task, constraints, and output format.
System prompt
Standing instructions that apply to every message in a conversation, not just one.
Context window
How much text a model can hold at once. Past the limit, the earliest material stops mattering.
Workflow
A fixed sequence of steps you designed. Predictable, cheap, easy to debug.
Agent
A model that decides its own next step in a loop, using tools, until a goal is met.
Tool / function call
A capability you hand the model — search, database write, send message — that it may choose to use.
RAG
Retrieval-augmented generation: fetch the relevant documents first, then answer from them.
Webhook
A URL that another system calls when something happens. The usual entry point of an automation.
Guardrail
A rule that constrains what the system may do or say, checked before the output reaches anyone.
Human in the loop
A required approval step before an automated action becomes visible to the outside world.
Hallucination
A fluent, confident, wrong answer. The model has no way to tell you it is guessing.