FredYakumo/zihuan-next
FredYakumo/zihuan-next has 41 GitHub stars and 4 forks. GitHot measured +1 stars in 24 hours and +1 in 7 days. Last checked 2026-11-11 14:44 UTC. Last push 2026-11-05.
Zihuan is a high-performance, highly customizable Rust-based AI Agent platform. Zihuan Impls a interesting Workspace Agent (zihuan code for coding) and an IM Agent (QQ bots, etc.)
Measured star growth
- in 24h
- +1
- in 7 days
- +1
- in 15 days
- Update delayed
- in 1 month
- Update delayed
Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.
README.md
Original
ZiHuan Next
Zihuan is a high-performance, highly customizable Rust-based AI Agent platform. Zihuan Impls a interesting Workspace Agent (zihuan code for coding) and an IM Agent (QQ bots, etc.) (I think these two cover pretty much all scenarios, but I'll add more if new use cases come up) as the basic framework. And Zihuan can easily tweak agent behavior and model inference details—whether using Candle, llama.cpp, or online APIs in various formats.
紫幻是一个高性能和高可定制化的Rust的AI Agent平台,现在紫幻实现了一套我觉得比较有趣的Workspace Agent(zihuan code,写代码的)和即时通讯软件Agent(QQ机器人啥的) (我觉得目前这两种Agent场景已经覆盖完全了,当然如果以后还有新的方式,也会支持) 作为基础能力框架. 然后,紫幻还可以更容易的定制Agent运行的行为细节,和Agent的大脑-模型 推理运行的细节(可以使用多种多样的模型,无论是使用Candle或者llama.cpp,还是各种格式的在线API)。
你可以参考文档查看Agent细节。
Quick Start
下载
解压后运行zihuan_next
- 不需要使用 Zihuan 运行本地模型时,可以选择 CPU 版本
- CUDA 版本仅适用于 NVIDIA GPU,并且需要安装与 CUDA 12.6 兼容的运行时依赖
- 其他架构或历史版本: Release。
紫幻启动的时候会在用户数据目录创建或者读取紫幻运行时的相关配置文件,你至少需要一个基础Agent来让你蹬紫幻,所以紫幻首次运行时还会引导你安装和配置一个Agent。
Features
聊天
I think the most interesting part of designing a bot is giving it long-term memory and personal likes/dislikes.
It can't like Millet today and then like chrysanthemum tomorrow, nor can it forget the people it has interacted with. Those bots that just call an LLM (plus various tools) to spew out tons of output are obviously boring.
While Zihuan supports all the tools you can think of, or tools you develop for it, it also tries to remember every day.
我觉得设计一个机器人最有趣的事情还是让它有自己的长期记忆和偏好好恶,
它不能今天喜欢某谷物公司明天又喜欢某为,也不能记不住跟它产生交集的人,那种只会调用LLM(然后加上各种tool)产生一大堆输出的机器人显然很无聊。
紫幻在支持了各种你能想到的工具,或者你为它开发能力工具的同时,它还会尝试记住每天。
Contribute
welcome your feedback on usage, bug reports, or suggestions for new features in the Issues.
If you find Zihuan is interesting, fun, you are also welcome to join our development team and contribute code in the PR.
欢迎在Issues里提出使用意见,Bug反馈,或者如果你希望加入什么新功能等。
如果你觉得紫幻很有趣很好玩,也欢迎你加入一起开发,在PR里贡献代码。
Contributors
Thanks to everyone who has contributed to Zihuan Next!
License
AGPL-3.0. See LICENSE.
Explore related projects
tensorflow/tensorflow
An Open Source Machine Learning Framework for Everyone
201,505 starshuggingface/transformers
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
168,815 starsAUTOMATIC1111/stable-diffusion-webui
Stable Diffusion web UI
165,289 starsComfy-Org/ComfyUI
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
146,271 stars