# AutoGenesisAgent: Self-Generating Multi-Agent Systems for Complex Tasks

arXiv: https://arxiv.org/abs/2404.17017 · HTML: https://mlsanity.com/paper/2404.17017/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 5 of its 10 references that are in the map.
- First posted on arXiv 25 Apr 2024.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 5 of 5)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [The Future of Scientific Publishing: Automated Article Generation](https://mlsanity.com/paper/2404.17586/) (2024)
- [Automated Extraction and Maturity Analysis of Open Source Clinical Informatics Repositories from Scientific Literature](https://mlsanity.com/paper/2403.14721/) (2024)

## Built on by (most cited first, 5 of 5)

- [Alita-G: Self-Evolving Generative Agent for Agent Generation](https://mlsanity.com/paper/2510.23601/) (2025)
- [A Misleading Gallery of Fluid Motion by Generative Artificial Intelligence](https://mlsanity.com/paper/2405.15406/) (2024)
- [Empowering Real-World: A Survey on the Technology, Practice, and Evaluation of LLM-driven Industry Agents](https://mlsanity.com/paper/2510.17491/) (2025)
- [SOCIA-nabla: Textual Gradient Meets Multi-Agent Orchestration for Automated Simulator Generation](https://mlsanity.com/paper/2505.12006/) (2025)
- [SOCIA-Nabla: Textual Gradient Meets Multi-Agent Orchestration for Automated Simulator Generation](https://mlsanity.com/paper/2510.18551/) (2025)

## Abstract (arXiv)

The proliferation of large language models (LLMs) and their integration into multi-agent systems has paved the way for sophisticated automation in various domains. This paper introduces AutoGenesisAgent, a multi-agent system that autonomously designs and deploys other multi-agent systems tailored for specific tasks. AutoGenesisAgent comprises several specialized agents including System Understanding, System Design, Agent Generator, and several others that collectively manage the lifecycle of creating functional multi-agent systems from initial concept to deployment. Each agent in AutoGenesisAgent has distinct responsibilities ranging from interpreting input prompts to optimizing system performance, culminating, in the deployment of a ready-to-use system. This proof-of-concept study discusses the design, implementation, and lessons learned from developing AutoGenesisAgent, highlighting its capability to generate and refine multi-agent systems autonomously, thereby reducing the need for extensive human oversight in the initial stages of system design. Keywords: multi-agent systems, large language models, system design automation, agent architecture, autonomous systems, software deployment