AI Agent Frameworks
AI Agent Frameworks is a hands-on playground and comparison of modern open-source AI agent and multi-agent frameworks, with practical examples for each one ranging from basic tasks to production-ready setups.
AI Agent Frameworks organizes practical examples for a wide range of agent frameworks into one repository, including AG2, Agno, AutoGen, the Claude Agent SDK, CrewAI, Deep Agents, Google ADK, LangChain, LangGraph, LlamaIndex, Microsoft Agent Framework, the OpenAI Agents SDK, Pydantic AI, smolagents and the Strands Agents SDK. Examples for each framework range from simple agent tasks to multi-agent workflows, retrieval-augmented generation, API integrations, and protocols such as A2A and MCP, with a dedicated README per framework folder.
A separate comparison module provides unified agent interfaces across several of the frameworks (Agno, LangGraph, LlamaIndex, OpenAI and Pydantic AI) and runs performance benchmarks measuring response time, token usage and tool utilization on common tasks, to compare frameworks on the same basis rather than relying on separate isolated examples.
Each framework folder is managed independently with uv for dependencies, and the project is written in Python.