OS-Copilot: Towards Generalist Computer Agents with Self-Improvement

🔥 News

• 2024.9: 🎉 Now Friday is equipped with vision! Try out the new friday_vision! Currently still under development but more stable versions are expected soon.
• 2024.6: 🎉 The front-end interface of OS-Copilot is now available. Go check it out in the frontend directory!
• 2024.3: 🎉 OS-Copilot is accepted at the LLM Agents Workshop@ICLR 2024!

What is OS-Copilot

OS-Copilot is an open-source library to build generalist agents capable of automatically interfacing with comprehensive elements in an operating system (OS), including the web, code terminals, files, multimedia, and various third-party applications.

⚡️ Quickstart

1. Clone the GitHub Repository:

   git clone https://github.com/OS-Copilot/OS-Copilot.git
   

2. Set Up Python Environment and Install Dependencies:

   conda create -n oscopilot_env python=3.10 -y
   conda activate oscopilot_env
   
   cd OS-Copilot
   pip install -e .
   

3. Set OpenAI API Key: Configure your OpenAI API key in .env.

   cp .env_template .env
   

4. Now you are ready to have fun:

   python quick_start.py
   

* FRIDAY currently only supports single-round conversation.

🛠️ Tutorial

| Level | Tutorial | Description | | ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------ | | Beginner | Installation | Explore three methods to install FRIDAY. | | Beginner | Getting Started | The simplest demonstration of FRIDAY with a quick_start.py script. | | Intermediate | Adding Your Tools | Adding and removing tools to the FRIDAY. | | Intermediate | Deploying API Services | Demonstrate the deployment of API services for FRIDAY. | | Intermediate | Example: Automating Excel Tasks | Automating Excel control using FRIDAY. | | Intermediate | Enhancing FRIDAY with Self-Learning for Excel Task Automation | Improved Excel control with self-directed learning. | | Advanced | Designing New API Tools | Guides on deploying custom API tools for FRIDAY to extend its functionalities. |

💻 User Interface (UI)

Enhance Your Experience with Our Intuitive Frontend! This interface is crafted for effortless control of your agents. For more details, visit OS-Copilot Frontend.

🏫 Community

Join our community to connect with other enthusiasts, researchers and developers:

• Discord: Join our Discord server for real-time discussions and support.
• Twitter: Follow our Twitter to get latest new or tag us to share your demos!

👨‍💻‍ Contributing

Visit the roadmap to preview what the community is working on and become a contributor!

🛡 Disclaimer

OS-Copilot is provided "as is" without warranty of any kind. Users assume full responsibility for any risks associated with its use, including potential data loss or changes to system settings. The developers of OS-Copilot are not liable for any damages or losses resulting from its use. Users must ensure their actions comply with applicable laws and regulations.

🔎 Citation

@article{wu2024copilot,
  title={Os-copilot: Towards generalist computer agents with self-improvement},
  author={Wu, Zhiyong and Han, Chengcheng and Ding, Zichen and Weng, Zhenmin and Liu, Zhoumianze and Yao, Shunyu and Yu, Tao and Kong, Lingpeng},
  journal={arXiv preprint arXiv:2402.07456},
  year={2024}
}

📬 Contact

If you have any inquiries, suggestions, or wish to contact us for any reason, we warmly invite you to email us at wuzhiyong@pjlab.org.cn.

Star History