This CrewAI workflow creates a 3-agent crew: a Researcher that browses the web for information, an Analyst that synthesizes findings and identifies gaps, and a Writer that produces a polished report. The agents work in sequence with shared memory.
TL;DR — Key Takeaways
- Multi-Agent Research Team with CrewAI — Intermediate level
- Tool: CrewAI
- 5 steps with detailed instructions
- Prerequisites: Python 3.10+, CrewAI installed, OpenAI API key, optional search API key (SerperDev)
Use Case
Research teams, content creators, and analysts who need comprehensive research reports on any topic.
Prerequisites
Python 3.10+, CrewAI installed, OpenAI API key, optional search API key (SerperDev)
Workflow Diagram
Multi-Agent Research Team with CrewAI
Step-by-Step Guide
-
Install CrewAI
pip install crewai. Create a new project. Define your agents in a crew.py file. CrewAI uses a simple YAML or Python config for agent definition.
-
Define the Researcher Agent
Create an agent with role: Researcher, goal: Find comprehensive information about the topic, backstory: Expert researcher with 10 years experience. Assign it a search tool (SerperDev, Google, or DuckDuckGo).
-
Define the Analyst Agent
Role: Analyst, goal: Synthesize research findings and identify key insights and gaps, backstory: Data analyst who excels at finding patterns. This agent receives the Researcher output and processes it.
-
Define the Writer Agent
Role: Writer, goal: Produce a clear, well-structured report from the analysis, backstory: Professional writer with expertise in making complex topics accessible. This agent receives the Analyst output.
-
Run the Crew
crew.kickoff(inputs={topic: your topic}). The agents execute in order: Researcher searches, Analyst processes, Writer formats. Review the final output and iterate on agent definitions for better results.
Tags
This workflow answers:
- How to set up multi-agent research team with crewai
- What tools and access you need
- How to troubleshoot common issues