CrewAI 18 Drills: Build Multi-Agent Crews & Flows on Gemini

About this course

This course contains the use of artificial intelligence."

Ask one AI assistant to do everything, and the instructions grow so long that quality falls apart" — this course exists to get you past that wall with a team that splits the work by role.

CrewAI is a framework that gives multiple AI agents a role, a goal, and a backstory, then makes them cooperate as a team (a Crew). And in 1.x, alongside Crews that cooperate autonomously, you can build Flows: business workflows with state management and conditional branching. This course builds up those two pillars, Crews and Flows, hands on, through 18 practical drills.

Learning follows a roadmap of 7 chapters. Once Section 0 has your environment ready, Section 1 nails the foundation of the three elements—Agent, Task, and Crew—how role/goal/backstory become the prompt itself, how expected_output decides output quality, and how you pass variables to kickoff with inputs.

Section 2 is the peak of the first half: coordinating multiple agents. You feel why roles get split with a two-agent Crew of a researcher and a writer, and watch each stage's result flow to the next through context between Tasks. Then you add an editor for a three-agent Crew to see how the output improves, and save the result as Markdown with output_file. Section 3 connects agents to the outside world. You read a local file with a built-in crewai-tools tool, write your own tool with the @tool decorator, and assemble a real-world-style Crew that reads rough notes and turns them into a formatted report.

In Section 4 you move definitions that were hard-coded into YAML configuration and a @CrewBase class—the shape of code you actually read in real projects and the official docs. Then Section 5 is the peak of the second half: Flows. You chain steps with @start and @listen, manage state as typed data with a Pydantic model, branch on an LLM's judgment with @router, and finally call a Crew from a Flow—reaching the

Prices, discounts and availability are set by Udemy. We may earn a commission when you purchase through links on this site.