Agentic Frameworks Lab

A comparative study of 6 popular AI agentic frameworks implementing identical use cases side-by-side.

Frameworks Included

• AutoGPT (v0.6.41)
• CrewAI (v1.8.1)
• Google ADK (v1.22.1)
• LangChain (v1.2.6) + LangGraph (v1.0.7)
• LlamaIndex (v0.14.13)
• Microsoft Agent Framework

Framework Comparisons

Comparison	Description
01-llm-tool-calling	Basic agent with tool execution (weather lookup)
02-multi-agent-orchestration	Multi-agent travel planning with role specialization
03-rag-implementation	Product Q&A with FAISS vector database
04-memory-management	Shopping assistant with persistent memory

Quick Start

1. Clone and setup

   cd agentic-frameworks-lab
   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Configure API keys

   cp .env.example .env
   # Edit .env and add your API keys

Step-by-Step Guide

Comparison 01: LLM Tool Calling

Use Case: Basic agent that can call external tools (weather lookup). Demonstrates fundamental tool execution patterns.

Run any framework:

# AutoGPT
cd framework-comparisons/01-llm-tool-calling/autogpt
pip install -r requirements.txt
python weather_agent.py

# CrewAI
cd framework-comparisons/01-llm-tool-calling/crewai
pip install -r requirements.txt
python weather_agent.py

# Google ADK
cd framework-comparisons/01-llm-tool-calling/google-adk
pip install -r requirements.txt
python weather_agent.py

# LangChain
cd framework-comparisons/01-llm-tool-calling/langchain
pip install -r requirements.txt
python weather_agent.py

# LlamaIndex
cd framework-comparisons/01-llm-tool-calling/llamaindex
pip install -r requirements.txt
python weather_agent.py

# Microsoft Agent Framework
cd framework-comparisons/01-llm-tool-calling/microsoft-agent-framework
pip install -r requirements.txt
python weather_agent.py

Comparison 02: Multi-Agent Orchestration

Use Case: Travel planning with 3 specialized agents (Researcher, Booking Specialist, Itinerary Planner) working together. Demonstrates role-based coordination and sequential task execution.

Run any framework:

# AutoGPT
cd framework-comparisons/02-multi-agent-orchestration/autogpt
pip install -r requirements.txt
python travel_planner.py

# CrewAI
cd framework-comparisons/02-multi-agent-orchestration/crewai
pip install -r requirements.txt
python travel_planner.py

# Google ADK
cd framework-comparisons/02-multi-agent-orchestration/google-adk
pip install -r requirements.txt
python travel_planner.py

# LangChain
cd framework-comparisons/02-multi-agent-orchestration/langchain
pip install -r requirements.txt
python travel_planner.py

# LlamaIndex
cd framework-comparisons/02-multi-agent-orchestration/llamaindex
pip install -r requirements.txt
python travel_planner.py

# Microsoft Agent Framework
cd framework-comparisons/02-multi-agent-orchestration/microsoft-agent-framework
pip install -r requirements.txt
python travel_planner.py

Comparison 03: RAG Implementation

Use Case: Product Q&A using a FAISS vector database with phone specifications. Demonstrates retrieval-augmented generation for knowledge-based responses.

Step 1 - Build the shared FAISS index (required once):

cd framework-comparisons/03-rag-implementation
pip install -r requirements.txt
python build_index.py

Step 2 - Run any framework:

# AutoGPT
cd framework-comparisons/03-rag-implementation/autogpt
pip install -r requirements.txt
python product_qa.py

# CrewAI
cd framework-comparisons/03-rag-implementation/crewai
pip install -r requirements.txt
python product_qa.py

# Google ADK
cd framework-comparisons/03-rag-implementation/google-adk
pip install -r requirements.txt
python product_qa.py

# LangChain
cd framework-comparisons/03-rag-implementation/langchain
pip install -r requirements.txt
python product_qa.py

# LlamaIndex
cd framework-comparisons/03-rag-implementation/llamaindex
pip install -r requirements.txt
python product_qa.py

# Microsoft Agent Framework
cd framework-comparisons/03-rag-implementation/microsoft-agent-framework
pip install -r requirements.txt
python product_qa.py

Comparison 04: Memory Management

Use Case: Shopping assistant with persistent memory across conversations. Implements short-term (conversation buffer), long-term (JSON file), and summarization patterns.

Run any framework:

# AutoGPT
cd framework-comparisons/04-memory-management/autogpt
pip install -r requirements.txt
python shopping_assistant.py

# CrewAI
cd framework-comparisons/04-memory-management/crewai
pip install -r requirements.txt
python shopping_assistant.py

# Google ADK
cd framework-comparisons/04-memory-management/google-adk
pip install -r requirements.txt
python shopping_assistant.py

# LangChain
cd framework-comparisons/04-memory-management/langchain
pip install -r requirements.txt
python shopping_assistant.py

# LlamaIndex
cd framework-comparisons/04-memory-management/llamaindex
pip install -r requirements.txt
python shopping_assistant.py

# Microsoft Agent Framework
cd framework-comparisons/04-memory-management/microsoft-agent-framework
pip install -r requirements.txt
python shopping_assistant.py

Project Structure

agentic-frameworks-lab/
├── framework-comparisons/
│   ├── 01-llm-tool-calling/
│   ├── 02-multi-agent-orchestration/
│   ├── 03-rag-implementation/
│   └── 04-memory-management/
│       ├── autogpt/
│       ├── crewai/
│       ├── google-adk/
│       ├── langchain/
│       ├── llamaindex/
│       └── microsoft-agent-framework/
├── .env.example
└── README.md

Requirements

• Python 3.x
• OpenAI API key (primary)
• Google Gemini API key (required for Google ADK examples)