CrewAI Multi-Agent Orchestration

Build teams of autonomous AI agents that collaborate on complex tasks.

When to Use

• Building multi-agent systems with specialized roles

• Need autonomous collaboration between agents

• Want role-based task delegation (researcher, writer, analyst)

• Require sequential or hierarchical process execution

• Building production workflows with memory and observability

Quick Start

pip install crewai pip install 'crewai[tools]' # With 50+ built-in tools

Simple Crew

from crewai import Agent, Task, Crew, Process

Define agents

researcher = Agent( role="Senior Research Analyst", goal="Discover cutting-edge developments in AI", backstory="You are an expert analyst with a keen eye for trends.", verbose=True )

writer = Agent( role="Technical Writer", goal="Create clear, engaging content about technical topics", backstory="You excel at explaining complex concepts.", verbose=True )

Define tasks

research_task = Task( description="Research the latest developments in {topic}. Find 5 key trends.", expected_output="A detailed report with 5 bullet points.", agent=researcher )

write_task = Task( description="Write a blog post based on the research findings.", expected_output="A 500-word blog post in markdown format.", agent=writer, context=[research_task] # Uses research output )

Create and run crew

crew = Crew( agents=[researcher, writer], tasks=[research_task, write_task], process=Process.sequential, verbose=True )

result = crew.kickoff(inputs={"topic": "AI Agents"}) print(result.raw)

Process Types

Sequential

Tasks execute in order:

crew = Crew( agents=[researcher, writer], tasks=[research_task, write_task], process=Process.sequential # Task 1 -> Task 2 -> Task 3 )

Hierarchical

Auto-creates a manager agent that delegates:

crew = Crew( agents=[researcher, writer, analyst], tasks=[research_task, write_task, analyze_task], process=Process.hierarchical, manager_llm="gpt-4o" )

Using Tools

from crewai_tools import ( SerperDevTool, # Web search ScrapeWebsiteTool, # Web scraping FileReadTool, # Read files PDFSearchTool, # Search PDFs )

researcher = Agent( role="Researcher", goal="Find accurate information", backstory="Expert at finding data online.", tools=[SerperDevTool(), ScrapeWebsiteTool()] )

Custom Tools

from crewai.tools import BaseTool

class CalculatorTool(BaseTool): name: str = "Calculator" description: str = "Performs mathematical calculations."

def _run(self, expression: str) -> str:
    try:
        return f"Result: {eval(expression)}"
    except Exception as e:
        return f"Error: {str(e)}"

YAML Configuration

agents.yaml

researcher: role: "{topic} Senior Data Researcher" goal: "Uncover cutting-edge developments in {topic}" backstory: > You're a seasoned researcher with a knack for uncovering the latest developments in {topic}.

reporting_analyst: role: "Reporting Analyst" goal: "Create detailed reports based on research data" backstory: > You're a meticulous analyst who transforms raw data into actionable insights.

tasks.yaml

research_task: description: > Conduct thorough research about {topic}. Find the most relevant information for {year}. expected_output: > A list with 10 bullet points of the most relevant information about {topic}. agent: researcher

reporting_task: description: > Review the research and create a comprehensive report. expected_output: > A detailed report in markdown format. agent: reporting_analyst output_file: report.md

Memory System

crew = Crew( agents=[researcher], tasks=[research_task], memory=True, # Enable memory embedder={ "provider": "openai", "config": {"model": "text-embedding-3-small"} } )

LLM Providers

from crewai import LLM

llm = LLM(model="gpt-4o") # OpenAI llm = LLM(model="claude-sonnet-4-5-20250929") # Anthropic llm = LLM(model="ollama/llama3.1", base_url="http://localhost:11434") # Local

agent = Agent(role="Analyst", goal="Analyze data", llm=llm)

Best Practices

• Clear roles - Each agent should have a distinct specialty

• YAML config - Better organization for larger projects

• Enable memory - Improves context across tasks

• Set max_iter - Prevent infinite loops (default 15)

• Limit tools - 3-5 tools per agent max

• Rate limiting - Set max_rpm to avoid API limits

Common Issues

Agent stuck in loop:

agent = Agent( role="...", max_iter=10, max_rpm=5 )

Task not using context:

task2 = Task( description="...", context=[task1], # Explicitly pass context agent=writer )

vs Alternatives

Feature CrewAI LangChain LangGraph

Best for Multi-agent teams General LLM apps Stateful workflows

Learning curve Low Medium Higher

Agent paradigm Role-based Tool-based Graph-based

Resources

• GitHub: https://github.com/crewAIInc/crewAI

• Docs: https://docs.crewai.com

• Tools: https://github.com/crewAIInc/crewAI-tools