嘿嘿,gezi 老师比我们 announcement 发的更早哈
Nowledge Mem 的原有的背景智能、知识处理、甚至最近引入的 Library(文档、笔记、URL)资料库,都和 LLM Wiki 不谋而合,老用户看到都是会心一笑的。
然而,LLM Wiki 本身是一个以 Wiki 为载体的工作流,它使得每一个知识单元的大小非常适合人 + Agent 的维护和消费,Wiki 中的双链特性让知识保有本应拥有的链接,这种形式在 Nowledge Mem 中还是缺失的。
最新版本里,我们重新审视 LLM Wiki 的工作流,结合我们 Mem 的天性、优势,把这个补上了,而且补得很自然,欢迎大家来体验、反馈。
Nowledge Mem comes with a builtin launcher for quickly search Memory/Insights from Mem in deep or fast mode.
But if you are already a Raycast Hacker(like us!🥰),
from today, we can install it from Raycast Store!
v0.6 upgrade in action → LanceDB now powers our dedicated text + vector search.
Result: hybrid (vector + BM25 + RRF), native multilingual support, ~50% faster deep search.
Grateful for the spotlight @lancedb
read the full story: https://t.co/T4H2eimSuc 🚀
Nowledge Mem v0.6 is released!
For the First Time in Forever, we brought most of what we imagined to the real world, the initial version already got some great feedbacks, we are keeping improving it and will bring more exciting features in our roadmap!