BuildingMemobase, a user profile-based memory backend.

📌 Beijing
👀 Large Language Model, Retrieve Augmented Generation(RAG)
🏃 Plan to start a project about addingMemory/State Layer for LLM
🏫 中国科学技术大学(USTC)
🏢 MSRA/Tencent/IDEA/沐言智语 -> memodb.io

🍠 小红书@坚白Gustavo, 我会定期分享RAG和LLM相关的一些projects
🐦 𝕏@坚白Gustavo, 随机想法
📧jianbaiye AT outlook DOT com

• You're interested in Open-Source and don't know how to start. 你对开源项目很感兴趣，但是不知道怎么开始。
• You're interested in LLM-native Applications. 你对LLM原生应用很感兴趣。
• You're interested in better RAG. 你对更好的RAG应用感兴趣。

• Flat is better than nested.
• Try not to make new stuffs.
• Printability is a great feature.
• Write something before you code.
• Overly abstract means narcissism.
• Hiding errors is just plain cocky.
• Long variable names doesn't hurt anyone.

PinnedLoading

1. memodb-io/memobasememodb-io/memobasePublic

   Profile-Based Long-Term Memory for AI Applications

   Python 951 61

2. nano-graphragnano-graphragPublic

   A simple, easy-to-hack GraphRAG implementation

   Python 2.7k 271

3. memodb-io/drive-flowmemodb-io/drive-flowPublic

   Build event-driven workflows with python async functions

   Python 33

4. nano-vectordbnano-vectordbPublic

   A simple, easy-to-hack Vector Database

   Python 130 6

5. chat-spotchat-spotPublic

   A Spotlight app. You can talk and snip anything to ChatGPT at your finger-tips

   TypeScript 36 8

6. LightGCN-PyTorchLightGCN-PyTorchPublic archive

   The PyTorch implementation of LightGCN

   Python 931 244