1. Project Overview

Zhiyi Cloud Nest

Zhiyi Cloud Nest is an information system or data collection for storing and organizing knowledge, used to save, manage, and access structured or unstructured information knowledge base.

This project is developed with Rui as the underlying framework.

2. Open Source License

Licensed under GPL-3.0

3. Screenshots

4. Feature Overview

Function List

First-level Function	Second-level Function	Version Number
Home Page		V1.0
Zhiyi Platform	Zhiyi Dialogue	V1.0
System Management	User Management	V1.0
	Role Management	V1.0
	Menu Management	V1.0
	Department Management	V1.0
	Position Management	V1.0
	Dictionary Management	V1.0
	Parameter Settings	V1.0
	Notifications and Announcements	V1.0
	Log Management	V1.0
System Monitoring	Online Users	V1.0
	Scheduled Tasks	V1.0
	Data Monitoring	V1.0
	Service Cache	V1.0
	Cache List	V1.0
System Tools	Form Builder	V1.0
	Code Generation	V1.0
	System Interface	V1.0
Document Management	Document Entry	V1.0
	Document Vectorization	V1.0

5. Tech Stack

Architecture Diagram

Technical Selection

Technical Type	Technical Name
Front-end Framework	Vue3
Back-end Architecture	Springboot
AI Framework	Langchain4j
Vector Database	Pgvector
Document Database	MongoDB
LLM Large Model	DeepSeek
Load Balancer	Nginx

 

LangChain Framework Introduction

LangChain is a framework for developing applications driven by language models. We believe that the most powerful and different applications will not only call language models through APIs, but also:

 

Data Awareness: Connecting language models with other data sources.

Proactivity: Allowing language models to interact with their environment.

Therefore, the design goal of the LangChain framework is to achieve these types of applications.

LangChain framework provides two main value propositions:

 

Components: LangChain provides modular abstracts for components needed to process language models. LangChain also provides a collection of implementations for all these abstractions. These components are intended to be easy to use, regardless of whether you use the rest of the LangChain framework.

Use-case-specific chains: Chains can be seen as a way of assembling these components in a specific way to best complete a specific use case. This is intended to be a higher-level interface that makes it easy to start with a specific use case. These chains are also intended to be customizable.

Therefore, we will divide this document into these two value propositions. In this document, we introduce components and use cases at a high level and in a language-agnostic way. For language-specific ways of using these components and solving these use cases, please see the language-specific part linked at the top of the page.

What is LangChain4j?

LangChain4j is the implementation of LangChain in the Java world.

It allows Java developers to easily access mainstream LLMs such as OpenAI, Gemini, Claude, and Mistral, while also encapsulating a suite of chain calls, tool integrations, and memory management functions.

 

In one sentence: LangChain4j = Java + LLM Best Partner.

Supported LLM Large Models in LangChain4j

In the world of LangChain4j, you can interface with various mainstream large language models (LLMs), and their APIs are encapsulated in a consistent and smooth way.

Pgvector is an open-source vector search engine, in addition to possessing all the characteristics of the PostgreSQL database, its most notable feature is that it can store and retrieve vector data in the PostgreSQL database, supporting precise and fuzzy searches for vectors. Vector formats, in addition to traditional embedding models' single-precision floating-point numbers, also support half-precision floating-point numbers, binary vectors, or sparse vectors.