OpenAGI: Your Gateway to Creating AI Agents
OpenAGI is a comprehensive and user-friendly package designed for creating AI agents. It serves as an excellent tool for building agents within the AIOS platform, captivating both AI enthusiasts and developers alike with its versatility and simplicity.
Getting Started
New users will find it easy to begin with OpenAGI. The package can be installed using the Python Package Index (PyPI) with a simple pip command:
pip install pyopenagi
For those who prefer working with local installations, the following steps will guide you:
git clone https://agiresearch/OpenAGI
cd OpenAGI
pip install -e .
How to Use OpenAGI
Adding a New Agent
To create a new agent in OpenAGI, users must set up a specific folder structure under the pyopenagi/agents directory. This structure is crucial for organizing the main execution logic and configurations of the agent:
- pyopenagi/agents
- author
- agent_name
- agent.py # Main code for the agent execution logic
- config.json # Configuration settings for the agent
- meta_requirements.txt # Dependencies required for the agent
For those who wish to utilize external tools provided by OpenAGI, instructions are available in the external tools setup guide. If an agent requires new tools, they must be added to the pyopenagi/tools/ directory.
Uploading an Agent
Once an agent is developed and thoroughly tested, the creator can upload it to share with the community using the following command:
python pyopenagi/agents/interact.py --mode upload --agent <author_name/agent_name>
It's important to ensure the agent parameter matches the local folder name exactly.
Downloading an Agent
Interested in exploring other users' agent implementations? Agents can be downloaded using:
python pyopenagi/agents/interact.py --mode download --agent <author_name/agent_name>
Contributions
OpenAGI encourages community contributions. Detailed contribution guidelines are available in the CONTRIBUTE document. Users can engage with the project by addressing issues or submitting pull requests.
Research and Development
OpenAGI is also at the forefront of research. The open-source implementation related to the research paper "OpenAGI: When LLM Meets Domain Experts" can be explored within the project's research branch. This paper, presented at the NeurIPS conference in 2023, provides insightful exploration into the integration of OpenAGI with domain expertise.
@article{openagi,
title={OpenAGI: When LLM Meets Domain Experts},
author={Ge, Yingqiang and Hua, Wenyue and Mei, Kai and Ji, Jianchao and Tan, Juntao and Xu, Shuyuan and Li, Zelong and Zhang, Yongfeng},
journal={In Advances in Neural Information Processing Systems (NeurIPS)},
year={2023}
}
Community and Recognition
OpenAGI thrives with a vibrant contributor community that continues to foster developments and innovation. See the contributors' list and become part of this growing ecosystem.
To trace the project's progress, users can view its star history, reflecting its reception and influence within the AI community.
In conclusion, OpenAGI is a robust pathway for both novice and seasoned developers to explore, create, and contribute to AI agent development, with a commitment to fostering an open, collaborative, and innovative environment.