Skip to content

HKUDS/AutoAgent

BranchesTags

Repository files navigation

Logo

AutoAgent: Fully-Automated & Zero-Code
LLM Agent Framework

Credits Join our Slack community Join our Discord community
Check out the documentation Paper Evaluation Benchmark Score

Welcome to AutoAgent! AutoAgent is a Fully-Automated and highly Self-Developing framework that enables users to create and deploy LLM agents through Natural Language Alone.

✨Key Features

  • 🏆 Top Performer on the GAIA Benchmark
    AutoAgent has ranked the #1 spot among open-sourced methods, delivering comparable performance to OpenAI's Deep Research.

  • 📚 Agentic-RAG with Native Self-Managing Vector Database
    AutoAgent equipped with a native self-managing vector database, outperforms industry-leading solutions like LangChain.

  • ✨ Agent and Workflow Create with Ease
    AutoAgent leverages natural language to effortlessly build ready-to-use tools, agents and workflows - no coding required.

  • 🌐 Universal LLM Support
    AutoAgent seamlessly integrates with A Wide Range of LLMs (e.g., OpenAI, Anthropic, Deepseek, vLLM, Grok, Huggingface ...)

  • 🔀 Flexible Interaction
    Benefit from support for both function-calling and ReAct interaction modes.

  • 🤖 Dynamic, Extensible, Lightweight
    AutoAgent is your Personal AI Assistant, designed to be dynamic, extensible, customized, and lightweight.

🚀 Unlock the Future of LLM Agents. Try 🔥AutoAgent🔥 Now!

Logo Quick Overview of AutoAgent.

🔥 News

  • [2025, Feb 17]:  🎉🎉We've updated and released AutoAgent v0.2.0 (formerly known as MetaChain). Detailed changes include: 1) fix the bug of different LLM providers from issues; 2) add automatic installation of AutoAgent in the container environment according to issues; 3) add more easy-to-use commands for the CLI mode. 4) Rename the project to AutoAgent for better understanding.
  • [2025, Feb 10]:  🎉🎉We've released MetaChain!, including framework, evaluation codes and CLI mode! Check our paper for more details.

📑 Table of Contents

🔍 How to Use AutoAgent

1. user mode (SOTA 🏆 Open Deep Research)

AutoAgent have an out-of-the-box multi-agent system, which you could choose user mode in the start page to use it. This multi-agent system is a general AI assistant, having the same functionality with OpenAI's Deep Research and the comparable performance with it in GAIA benchmark.

  • 🚀 High Performance: Matches Deep Research using Claude 3.5 rather than OpenAI's o3 model.
  • 🔄 Model Flexibility: Compatible with any LLM (including Deepseek-R1, Grok, Gemini, etc.)
  • 💰 Cost-Effective: Open-source alternative to Deep Research's $200/month subscription
  • 🎯 User-Friendly: Easy-to-deploy CLI interface for seamless interaction
  • 📁 File Support: Handles file uploads for enhanced data interaction
Input
Input your request. 		Output
Agent will give you the response. 		Select Agent
Use @ to mention the agent you want to use. (Optional)
Input
@Upload_files will help you upload the files. 		Output
Select the files you want to use. 		Select Agent
Successfully uploaded the files.

2. agent editor (Agent Creation without Workflow)

The most distinctive feature of AutoAgent is its natural language customization capability. Unlike other agent frameworks, AutoAgent allows you to create tools, agents, and workflows using natural language alone. Simply choose agent editor or workflow editor mode to start your journey of building agents through conversations.

You can use agent editor as shown in the following figure.

requirement
Input what kind of agent you want to create. 		profiling
Automated agent profiling. 		profiles
Output the agent profiles.
tools
Create the desired tools. 		task
Input what do you want to complete with the agent. (Optional) 		output
Create the desired agent(s) and go to the next step.

3. workflow editor (Agent Creation with Workflow)

You can also create the agent workflows using natural language description with the workflow editor mode, as shown in the following figure. (Tips: this mode does not support tool creation temporarily.)

requirement
Input what kind of workflow you want to create. 		profiling
Automated workflow profiling. 		profiles
Output the workflow profiles.
task
Input what do you want to complete with the workflow. (Optional) 		output
Create the desired workflow(s) and go to the next step.

⚡ Quick Start

Installation

AutoAgent Installation

git clone https://github.com/HKUDS/AutoAgent.git
cd AutoAgent
pip install -e .

Docker Installation

We use Docker to containerize the agent-interactive environment. So please install Docker first. You don't need to manually pull the pre-built image, because we have let Auto-Deep-Research automatically pull the pre-built image based on your architecture of your machine.

API Keys Setup

Create an environment variable file, just like .env.template, and set the API keys for the LLMs you want to use. Not every LLM API Key is required, use what you need.

# Required Github Tokens of your own
GITHUB_AI_TOKEN=

# Optional API Keys
OPENAI_API_KEY=
DEEPSEEK_API_KEY=
ANTHROPIC_API_KEY=
GEMINI_API_KEY=
HUGGINGFACE_API_KEY=
GROQ_API_KEY=
XAI_API_KEY=

Start with CLI Mode

[🚨 News: ] We have updated a more easy-to-use command to start the CLI mode and fix the bug of different LLM providers from issues. You can follow the following steps to start the CLI mode with different LLM providers with much less configuration.

Command Options:

You can run auto main to start full part of AutoAgent, including user mode, agent editor and workflow editor. Btw, you can also run auto deep-research to start more lightweight user mode, just like the Auto-Deep-Research project. Some configuration of this command is shown below.

  • --container_name: Name of the Docker container (default: 'deepresearch')
  • --port: Port for the container (default: 12346)
  • COMPLETION_MODEL: Specify the LLM model to use, you should follow the name of Litellm to set the model name. (Default: claude-3-5-sonnet-20241022)
  • DEBUG: Enable debug mode for detailed logs (default: False)
  • API_BASE_URL: The base URL for the LLM provider (default: None)
  • FN_CALL: Enable function calling (default: None). Most of time, you could ignore this option because we have already set the default value based on the model name.
  • git_clone: Clone the AutoAgent repository to the local environment (only support with the auto main command, default: True)
  • test_pull_name: The name of the test pull. (only support with the auto main command, default: 'autoagent_mirror')

More details about git_clone and test_pull_name]

In the agent editor and workflow editor mode, we should clone a mirror of the AutoAgent repository to the local agent-interactive environment and let our AutoAgent automatically update the AutoAgent itself, such as creating new tools, agents and workflows. So if you want to use the agent editor and workflow editor mode, you should set the git_clone to True and set the test_pull_name to 'autoagent_mirror' or other branches.

auto main with different LLM Providers

Then I will show you how to use the full part of AutoAgent with the auto main command and different LLM providers. If you want to use the auto deep-research command, you can refer to the Auto-Deep-Research project for more details.

Anthropic
  • set the ANTHROPIC_API_KEY in the .env file.
ANTHROPIC_API_KEY=your_anthropic_api_key
  • run the following command to start Auto-Deep-Research.
auto main # default model is claude-3-5-sonnet-20241022
OpenAI
  • set the OPENAI_API_KEY in the .env file.
OPENAI_API_KEY=your_openai_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=gpt-4o auto main
Mistral
  • set the MISTRAL_API_KEY in the .env file.
MISTRAL_API_KEY=your_mistral_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=mistral/mistral-large-2407 auto main
Gemini - Google AI Studio
  • set the GEMINI_API_KEY in the .env file.
GEMINI_API_KEY=your_gemini_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=gemini/gemini-2.0-flash auto main
Huggingface
  • set the HUGGINGFACE_API_KEY in the .env file.
HUGGINGFACE_API_KEY=your_huggingface_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=huggingface/meta-llama/Llama-3.3-70B-Instruct auto main
Groq
  • set the GROQ_API_KEY in the .env file.
GROQ_API_KEY=your_groq_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=groq/deepseek-r1-distill-llama-70b auto main
OpenAI-Compatible Endpoints (e.g., Grok)
  • set the OPENAI_API_KEY in the .env file.
OPENAI_API_KEY=your_api_key_for_openai_compatible_endpoints
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=openai/grok-2-latest API_BASE_URL=https://api.x.ai/v1 auto main
OpenRouter (e.g., DeepSeek-R1)

We recommend using OpenRouter as LLM provider of DeepSeek-R1 temporarily. Because official API of DeepSeek-R1 can not be used efficiently.

  • set the OPENROUTER_API_KEY in the .env file.
OPENROUTER_API_KEY=your_openrouter_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=openrouter/deepseek/deepseek-r1 auto main
DeepSeek
  • set the DEEPSEEK_API_KEY in the .env file.
DEEPSEEK_API_KEY=your_deepseek_api_key
  • run the following command to start Auto-Deep-Research.
COMPLETION_MODEL=deepseek/deepseek-chat auto main

After the CLI mode is started, you can see the start page of AutoAgent:

Logo Start Page of AutoAgent.

Tips

Import browser cookies to browser environment

You can import the browser cookies to the browser environment to let the agent better access some specific websites. For more details, please refer to the cookies folder.

Add your own API keys for third-party Tool Platforms

If you want to create tools from the third-party tool platforms, such as RapidAPI, you should subscribe tools from the platform and add your own API keys by running process_tool_docs.py.

python process_tool_docs.py

More features coming soon! 🚀 Web GUI interface under development.

☑️ Todo List

AutoAgent is continuously evolving! Here's what's coming:

  • 📊 More Benchmarks: Expanding evaluations to SWE-bench, WebArena, and more
  • 🖥️ GUI Agent: Supporting Computer-Use agents with GUI interaction
  • 🔧 Tool Platforms: Integration with more platforms like Composio
  • 🏗️ Code Sandboxes: Supporting additional environments like E2B
  • 🎨 Web Interface: Developing comprehensive GUI for better user experience

Have ideas or suggestions? Feel free to open an issue! Stay tuned for more exciting updates! 🚀

🔬 How To Reproduce the Results in the Paper

GAIA Benchmark

For the GAIA benchmark, you can run the following command to run the inference.

cd path/to/AutoAgent && sh evaluation/gaia/scripts/run_infer.sh

For the evaluation, you can run the following command.

cd path/to/AutoAgent && python evaluation/gaia/get_score.py

Agentic-RAG

For the Agentic-RAG task, you can run the following command to run the inference.

Step1. Turn to this page and download it. Save them to your datapath.

Step2. Run the following command to run the inference.

cd path/to/AutoAgent && sh evaluation/multihoprag/scripts/run_rag.sh

Step3. The result will be saved in the evaluation/multihoprag/result.json.

📖 Documentation

A more detailed documentation is coming soon 🚀, and we will update in the Documentation page.

🤝 Join the Community

We want to build a community for AutoAgent, and we welcome everyone to join us. You can join our community by:

🙏 Acknowledgements

Rome wasn't built in a day. AutoAgent stands on the shoulders of giants, and we are deeply grateful for the outstanding work that came before us. Our framework architecture draws inspiration from OpenAI Swarm, while our user mode's three-agent design benefits from Magentic-one's insights. We've also learned from OpenHands for documentation structure and many other excellent projects for agent-environment interaction design, among others. We express our sincere gratitude and respect to all these pioneering works that have been instrumental in shaping AutoAgent.

🌟 Cite

@misc{AutoAgent,
      title={{AutoAgent: A Fully-Automated and Zero-Code Framework for LLM Agents}},
      author={Jiabin Tang, Tianyu Fan, Chao Huang},
      year={2025},
      eprint={202502.05957},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2502.05957},
}

About

"AutoAgent: Fully-Automated and Zero-Code LLM Agent Framework"

arxiv.org/abs/2502.05957

Topics

agent llms

Resources

Readme

License

MIT license
Activity

Stars

325 stars

Watchers

11 watching

Forks

54 forks
Report repository

Releases

No releases published

Packages

No packages published

Contributors 4

Languages