Awesome AI Agents › AI Agent Operating System

HKUDS/AutoAgent

⭐ 9787 Python repository created 2025-02-06

AutoAgent is a fully-automated and zero-code framework designed to create and deploy large language model (LLM) agents using natural language alone. It offers a highly self-developing system that allows users to build tools, agents, and workflows without any coding. AutoAgent has achieved top performance on the GAIA benchmark, competing with OpenAI's Deep Research. It features an agentic retrieval-augmented generation (RAG) system with a native self-managing vector database that outperforms other solutions like LangChain. The framework supports a wide range of LLMs including OpenAI, Anthropic, Deepseek, vLLM, Grok, and Huggingface, providing flexibility in model choice. Interaction modes include function-calling and ReAct, enabling dynamic and flexible agent behavior. AutoAgent is designed to be lightweight, extensible, and customizable, serving as a personal AI assistant. It offers multiple usage modes such as user mode for out-of-the-box multi-agent systems, agent editor for creating agents without workflows, and workflow editor for building agent workflows, all through natural language instructions. The project includes CLI tools for easy deployment and supports file uploads for enhanced data interaction. It is an open-source alternative to costly subscription-based AI research assistants, aiming to democratize access to advanced LLM agent capabilities. The project is actively maintained with frequent updates and has a strong community presence on Slack, Discord, and Wechat. Comprehensive documentation and a research paper are available to support users and developers.

https://github.com/HKUDS/AutoAgent

agent-editoragentic-ragai-assistantanthropicautoagentclicustomizabledeepseekextensiblefully-automatedfunction-callinggaia-benchmarkgrokhuggingfacelangchainllm-agentmulti-agent-systemnatural-languageopen-sourceopenaipersonal-ai-assistantreactuser-modevector-databasevllmworkflow-editorzero-code