DOCSEARCHbyAlgolia
Submit Docs

JeeSite 5 Rapid Development Platform

repository·v5.springboot4·Indexed 23 days ago

https://github.com/thinkgem/jeesite5

An enterprise-grade, low-code rapid development platform based on Spring Boot. It features a micro-kernel architecture supporting SaaS, clustering, and microservices. Key capabilities include automated code generation, a Flowable workflow engine, and an AI module utilizing Spring AI and RAG technology. The AI module supports cloud and local models (Ollama, DeepSeek), vector databases (Chroma, PGVector, Elasticsearch, Milvus), and the Model Context Protocol (MCP) for tool calling.

Tokens
38.3K
Snippets
35
Records
208
Agent score
81%

What's inside JeeSite 5

  1. Overview of the JeeSite AI Module

    v5.springboot4

    The AI module is built on Spring AI and the JeeSite CMS, utilizing Retrieval-Augmented Generation (RAG) technology to create an enterprise-grade knowledge management and intelligent dialogue platform.

    Key features include:

    • RAG Technology: Automatically retrieves relevant information from enterprise documents (including attachments like Word, Excel, PPT, and ZIP) to provide accurate, context-aware answers.
    • Model Support: Supports both cloud-based models (via OpenAPI standard interfaces) and local models like Ollama, DeepSeek, and Tongyi Qianwen.
    • Vector Database Integration: Seamlessly integrates with vector stores such as Chroma, PGVector, Elasticsearch, and Milvus for efficient data storage and retrieval.
    • Extensibility: Supports Tool Calling (Function Calling) to link AI with your business logic and the Model Context Protocol (MCP) for server/client tool interaction.

  2. Overview of layer mobile

    v5.springboot4

    layer mobile is a specialized layer support library designed specifically for mobile devices (phones, tablets, and WebKit-based browsers/WebViews). It is written in Native JavaScript and is completely independent of the PC version of layer. You should choose between the PC and mobile versions based on your specific usage scenario.

    Key features include:

    • Zero dependencies: No external libraries are required; you only need to load layer.m.js.
    • High performance: Lightweight and optimized for mobile environments.
    • Adaptive design: Features robust adaptive capabilities for various mobile screen sizes.
    • Customizable skins: Supports flexible skin customization to match different UI styles.
    • Rich API: Provides a comprehensive set of interfaces for various popup/layer requirements.

  3. Overview of jeesite-web-ai-mcp

    v5.springboot4
    jeesite-web-ai-mcp is a Model Context Protocol (MCP) server for JeeSite. It enables external AI Large Language Models (LLMs) to interact with JeeSite services semantically. This allows an AI to perform tasks such as querying specific user information (e.g., "query user information for Zhang San") by calling JeeSite service data through the MCP interface.

  4. Overview of JeeSite 5 Rapid Development Platform

    v5.springboot4

    JeeSite is an enterprise-grade, low-code rapid development solution. It provides a comprehensive suite of tools for building complex business systems, including online data source management, data modeling, and automated code generation for both monolithic and microservice architectures.

    Key features include:

    • Core Modules: Organization, users, roles, permissions (menu/button/data), system monitoring, and dictionary management.
    • Extended Modules: Workflow engine (Flowable), content management, message pushing, SSO, third-party login, scheduled tasks, object storage, and data visualization.
    • AI Integration: Support for OpenAPI, Ollama, and DeepSeek using RAG (Retrieval-Augmented Generation) for enterprise knowledge bases, as well as MCP (Model Context Protocol) for tool calling.
    • Architecture: Decoupled, lightweight, micro-kernel design supporting SaaS, clustering, read-write splitting, and Spring Cloud microservices.

  5. Overview of jeesite-web-mini

    v5.springboot4

    jeesite-web-mini is a lightweight core service for JeeSite designed for extremely fast startup (approximately 1 second). It is intended for use as a backend internal data processing service.

    Key characteristics:

    • Does not initialize user interfaces or permission systems.
    • Can be configured to run without initializing a data source.
    • Optimized for microservices and small-scale backend tasks.

  6. Understand the difference between jeesite-web-api and jeesite-web

    v5.springboot4
    The jeesite-web-api project is the Web API component of JeeSite. Unlike jeesite-web, which is designed for traditional server-side rendering, jeesite-web-api is specifically engineered to provide interface services for decoupled/separated frontends (e.g., Single Page Applications). When using jeesite-web-api, the backend login pages and main framework UI pages are disabled to focus exclusively on API delivery.

  7. Understand the difference between jeesite-web-vue and jeesite-web

    v5.springboot4

    The jeesite-web-vue project is the Vue-based frontend implementation for JeeSite. Unlike the standard jeesite-web project, jeesite-web-vue is configured to:

    1. Default to a decoupled Vue interface.
    2. Disable the traditional HTML + jQuery-based login pages and main framework page displays.

  8. Understand JeeSite5 web module types

    v5.springboot4

    JeeSite5 provides several specialized web modules depending on your architecture needs:

    FolderDescription
    webFull-stack version using HTML5 + jQuery
    web-vueDecoupled frontend using Vue; dist runs in Tomcat
    web-apiAPI service specifically for decoupled Vue clients (no login UI)
    web-fastQuick experience project; uses H2 database; includes both Vue and HTML5 interfaces
    web-ai-cmsContains CMS and AI modules (AI Assistant, AI Chat, Tool MCP calls)
    web-ai-mcpMCP server for AI modules, providing Tool call implementations
    web-miniMinimal kernel service for internal data processing; fast startup (1s) and low resource usage

  9. Implement MCP (Model Context Protocol)

    v5.springboot4

    The module implements the Model Context Protocol (MCP) using Server-Sent Events (SSE).

    MCP Server Configuration

    Enable the MCP server and define the SSE endpoints:

    spring:
  ai:
    mcp:
      server:
        enabled: true
        sse-endpoint: /api/v1/sse
        sse-message-endpoint: /api/v1/mcp

    MCP Client Configuration

    Configure the client to connect to an MCP server:

    spring:
  ai:
    model:
      chat: ollama
    mcp:
      client:
        enabled: true
        sse:
          connections:
            jeesite:
              url: http://127.0.0.1:8992
              sse-endpoint: /api/v1/sse

  10. Enable Tool Calling (Function Calling)

    v5.springboot4

    Tool Calling allows the AI model to interact with your Java-implemented business logic. This enables the model to perform actions and integrate deeply with your existing services while maintaining user session and data permissions.

    To enable this feature, set the following property:

    spring.ai.tool-calls: true

    Refer to TestAiTools.java and UserAiTools.java in the source code for implementation examples.

  11. Clone the JeeSite5 repository

    v5.springboot4

    To obtain the source code, clone the repository from Gitee. Important: Do not execute these commands in a directory containing Chinese characters or spaces. After cloning, switch to the v5.springboot4 branch to use the Spring Boot 4 version.

    # Avoid directories with Chinese characters or spaces
git clone https://gitee.com/thinkgem/jeesite5.git jeesite
cd jeesite

# Switch to the Spring Boot 4 branch
git switch v5.springboot4
    git clone https://gitee.com/thinkgem/jeesite5.git jeesite
cd jeesite
git switch v5.springboot4