## Core Concepts ### Agent Definition ```python from crewai import Agent, Task, Crew, Process from langchain.tools import DuckDuckGoSearchRun researcher = Agent( role='Senior Research Analyst', goal='Uncover cutting-edge developments in AI and data science', backstory="""You work at a leading tech think tank. Your expertise lies in identifying emerging trends. You have a knack for dissecting complex data and presenting actionable insights.""", verbose=True, allow_delegation=False, tools=[DuckDuckGoSearchRun()], llm=llm ) writer = Agent( role='Tech Content Strategist', goal='Craft compelling content on tech advancements', backstory="""You are a renowned Content Strategist, known for your insightful and engaging articles. You transform complex concepts into compelling narratives.""", verbose=True, allow_delegation=True, llm=llm ) editor = Agent( role='Editor-in-Chief', goal='Ensure content quality and accuracy', backstory="""You are a meticulous editor with years of experience in tech journalism. You ensure all content meets high standards of quality and factual accuracy.""", verbose=True, allow_delegation=False, llm=llm ) ``` ### Task Definition ```python task1 = Task( description="""Conduct a comprehensive analysis of the latest advancements in AI as of 2024. Identify key trends, breakthrough technologies, and potential industry impacts. Your final answer MUST be a full analysis report""", agent=researcher, expected_output="A comprehensive 3-paragraph report on AI advancements" ) task2 = Task( description="""Using the research findings, write an engaging blog post that highlights the most significant AI advancements. Your post should be informative yet accessible.""", agent=writer, context=[task1], expected_output="A 4-paragraph blog post in markdown format" ) task3 = Task( description="""Review the blog post for clarity, accuracy, and engagement. Provide constructive feedback and suggest improvements.""", agent=editor, context=[task2], expected_output="A list of specific improvements and final approval" ) ``` ### Crew Assembly ```python content_team = Crew( agents=[researcher, writer, editor], tasks=[task1, task2, task3], verbose=2, process=Process.sequential, # Options: sequential, hierarchical memory=True, # Enable memory for context cache=True # Cache results ) result = content_team.kickoff() print(result) ``` ## Advanced Patterns ### Hierarchical Process ```python from crewai import Crew, Process content_team = Crew( agents=[researcher, writer, editor], tasks=[task1, task2, task3], process=Process.hierarchical, manager_llm=llm, # LLM for the manager agent verbose=2 ) ``` ### Custom Tools ```python from langchain.tools import tool @tool def analyze_sentiment(text: str) -> str: """Analyze sentiment of given text""" return "positive" if "good" in text else "negative" @tool def fetch_stock_data(ticker: str) -> dict: """Fetch stock data for ticker symbol""" return {"price": 150.0, "change": 2.5} analyst = Agent( role='Financial Analyst', tools=[analyze_sentiment, fetch_stock_data], ) ``` ### Memory and Context ```python from crewai.memory import EntityMemory research_crew = Crew( agents=[researcher, analyst], tasks=[task1, task2], memory=True, entity_memory=EntityMemory(), verbose=2 ) ``` ## Best Practices - Define clear, specific roles for each agent - Use allow_delegation wisely (only for managers) - Provide detailed backstories for better context - Chain tasks using context parameter - Enable memory for long-running workflows