From chatbots to agents: tool calling, MCP servers, frameworks like CrewAI, LangGraph, and AutoGen, memory systems, and running multiple agents in parallel. An agent without guardrails is basically an intern with root access, so you’ll also learn how to set the right boundaries.
Why this matters: Agents are the difference between AI that talks and AI that actually works.
The learning path below is organized in the order that makes the most sense: understand the concepts first, then try them out, then go deeper. Check off what you’ve already read or implemented. Your progress is saved locally in your browser.
Getting started and overview
AI agent fundamentals
Visualizing AI agents: bringing agents to life in an artificial island and office world
Agent workflows
An army of AI agents: why TikTok courses on agents are booming and what’s really behind it
AI agent frameworks
Running locally
MCP
Monitoring
Security
Done? Head over to the free AI learning path to continue with the next section.