# AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation Framework

**Authors:** Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Shaokun Zhang, Erkang Zhu, Beibin Li, Li Jiang, Xiaoyun Zhang, Chi Wang  
**arXiv:** [2308.08155](https://arxiv.org/abs/2308.08155)  
**Date:** 2023-08-16  
**Summary page:** https://www.summarizepaper.com/en/paper/2308.08155v1/

## Key points

- AutoGen is a framework for developing next-generation Language and Learning Models (LLMs) applications through multi-agent conversation.
- AutoGen agents are customizable, conversable, and integrate human participation.
- AutoGen agents operate in various modes, combining LLMs, human inputs, and tools to solve tasks.
- AutoGen simplifies complex LLM workflows as automated agent chats.
- Examples of applications that can be developed using AutoGen include coding problems, mathematics tasks, operations research challenges, entertainment applications, online decision-making scenarios, question answering systems, etc.
- The report discusses the limitations of using external tools like Google Scholar's unofficial API for paper search.
- Overall, AutoGen offers a promising framework for developing advanced LLM applications with enhanced automation and human interaction capabilities.

## Summary

This technical report introduces AutoGen, a novel framework that enables the development of next-generation Language and Learning Models (LLMs) applications through multi-agent conversation. AutoGen agents are customizable, conversable, and seamlessly integrate human participation to effectively navigate the imperfect generation and reasoning abilities of LLMs. They operate in various modes, combining LLMs, human inputs, and tools to solve tasks while leveraging human understanding and intelligence through agent conversations. The design of AutoGen simplifies complex LLM workflows as automated agent chats. The report provides diverse examples showcasing how developers can utilize AutoGen to solve coding problems, mathematics tasks, operations research challenges, entertainment applications, online decision-making scenarios, question answering systems, and more. Additionally, the authors discuss the limitations of using external tools like Google Scholar's unofficial API for paper search. Overall, AutoGen presents a promising framework for developing advanced LLM applications with enhanced automation and human interaction capabilities.