AI Agent Frameworks Comparison: CrewAI vs LangGraph vs AutoGen

This project demonstrates and compares three leading Python agent frameworks for orchestrating multi-agent financial analysis workflows:

• CrewAI: Role-based, sequential agent orchestration
• LangGraph: State-driven, graph-based agent workflows
• AutoGen: Conversational, group-chat style agent collaboration

All three frameworks are implemented to solve the same problem: analyzing a stock's recent performance and generating an investment recommendation (BUY/SELL/HOLD) with rationale.

---

📦 Project Structure

ai-agent-comparision/
├── crewai/      # CrewAI implementation
├── langgraph/   # LangGraph implementation
├── autogen/     # AutoGen implementation
├── requirements.txt
├── .env.example
└──  README.md

---

🚀 How to Use & Run Each Framework

1. Prerequisites

• Python 3.8+
• API keys for Groq and Tavily
• Install dependencies:

  python -m venv .venv
  source .venv/Scripts/activate  # Windows
  pip install -r requirements.txt
  cp .env.example .env  # and fill in your API keys

2. Running Each Implementation

CrewAI

cd crewai
python main.py

• Prompt: Enter a stock ticker (e.g. NVDA, AAPL)
• Output: Executive summary and recommendation

LangGraph

cd langgraph
python main.py

• Prompt: Enter a stock ticker
• Output: Analysis and recommendation

AutoGen

cd autogen
python main.py

• Prompt: Enter a stock ticker
• Output: Analyst/Researcher group chat, final report

---

🧩 Framework Comparison Table

Feature	CrewAI	LangGraph	AutoGen
Orchestration	Sequential pipeline	State graph (DAG)	Group chat (conversational)
Agent Roles	Explicit, role-based	Node-based, flexible	Conversational, flexible
Task Flow	Linear, step-by-step	Custom graph transitions	Multi-turn dialogue
Extensibility	Add agents/tasks	Add nodes/edges	Add agents, chat logic
Best For	Business workflows	Complex dependencies	Dynamic collaboration
Code Structure	agents.py, tasks.py, tools.py	nodes.py, state.py, tools.py	agents.py, workflow.py, config.py
Learning Curve	Low/Medium	Medium/High	Medium
Output	Executive report	Analysis + recommendation	Chat log + report

---

🛠️ How to Extend

• Add new data sources: Edit tools.py in any framework
• Change agent logic: Edit agents.py or nodes.py
• Add new analysis steps: Add new tasks/nodes/agents as appropriate

---

📚 Further Reading

• CrewAI Documentation
• LangGraph Documentation
• AutoGen Documentation

---

This project is a reference for anyone looking to build modular, multi-agent systems in Python using modern frameworks.