KidneyTalk-open

KidneyTalk-open is the open-source repository for the paper "KidneyTalk-open: No-code Deployment of a Private Large Language Model with Medical Documentation-Enhanced Knowledge Database for Kidney Disease".

This project aims to provide a no-code deployment system for private medical large language models, focusing on knowledge enhancement and clinical decision support in nephrology. The system integrates advanced open-source models and improves medical knowledge utilization through an adaptive retrieval augmentation framework. With its graphical interface design, clinical practitioners can conveniently manage medical documents and obtain AI-assisted decision support without programming skills.

Key Features:

• 🔒 Fully local deployment for patient privacy protection
• 📚 Support for integrating medical knowledge base with real-time retrieval augmentation
• 📋 Intelligent medical document processing
• 🖥️ Zero-barrier graphical interface
• 🔄 Adaptive knowledge retrieval framework

📦 Installation Package

You can now download the KidneyTalk-open installation package from our release page: https://github.com/PKUDigitalHealth/KidneyTalk-open/releases/tag/v1.0.0

We're excited to share this with the community and look forward to your feedback! 🎉

💻 Requirements

• MacOS: 13 or higher
• Ollama: Download and install Ollama from https://ollama.ai/. After installation, you can proceed with the next steps without worrying about Ollama anymore.

📺 Installation Tutorial Video

We've prepared a detailed video tutorial to guide you through the KidneyTalk installation process! 🎬

如果您无法访问Youtube，也可以在B站观看： KidneyTalk安装部署教程 ✨

Don't hesitate to follow along with the video - we've made sure to cover everything you need to get started! 💪

Acknowledgements

KidneyTalk-open builds on top of other open-source projects:

• Jan
• LangChain

License

KidneyTalk-open is free and open source, under the AGPLv3 license.

Citation

If you find KidneyTalk-open useful, please consider citing our work:

@misc{long2025kidneytalkopennocodedeploymentprivate,
      title={KidneyTalk-open: No-code Deployment of a Private Large Language Model with Medical Documentation-Enhanced Knowledge Database for Kidney Disease}, 
      author={Yongchao Long and Chao Yang and Gongzheng Tang and Jinwei Wang and Zhun Sui and Yuxi Zhou and Shenda Hong and Luxia Zhang},
      year={2025},
      eprint={2503.04153},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2503.04153}, 
}