- Explore MCP Servers
- zafu_xiaolin
Zafu Xiaolin
What is Zafu Xiaolin
zafu_xiaolin is a full-stack intelligent assistant solution designed for campus and educational scenarios, specifically developed for Zhejiang Agricultural and Forestry University. It serves as an AI assistant that integrates task planning and tool invocation, utilizing natural language as the medium for communication with large language models (LLMs).
Use cases
Use cases for zafu_xiaolin include providing students with personalized learning recommendations, accessing the latest campus notifications and event schedules, and ensuring secure handling of student data for privacy protection.
How to use
To use zafu_xiaolin, users can interact with the AI assistant through natural language queries. The assistant connects to various campus systems to fetch real-time information, provide personalized services, and assist with complex campus-related tasks. Users need to have the necessary prerequisites installed, such as Python, Node.js, and Redis.
Key features
Key features of zafu_xiaolin include seamless integration with campus data systems, real-time information retrieval, personalized service offerings based on student data, and secure data processing to protect sensitive information.
Where to use
zafu_xiaolin is primarily used in educational institutions, particularly in campus environments where it can connect with various systems like academic administration, library services, and student information systems.
Clients Supporting MCP
The following are the main client software that supports the Model Context Protocol. Click the link to visit the official website for more information.
Overview
What is Zafu Xiaolin
zafu_xiaolin is a full-stack intelligent assistant solution designed for campus and educational scenarios, specifically developed for Zhejiang Agricultural and Forestry University. It serves as an AI assistant that integrates task planning and tool invocation, utilizing natural language as the medium for communication with large language models (LLMs).
Use cases
Use cases for zafu_xiaolin include providing students with personalized learning recommendations, accessing the latest campus notifications and event schedules, and ensuring secure handling of student data for privacy protection.
How to use
To use zafu_xiaolin, users can interact with the AI assistant through natural language queries. The assistant connects to various campus systems to fetch real-time information, provide personalized services, and assist with complex campus-related tasks. Users need to have the necessary prerequisites installed, such as Python, Node.js, and Redis.
Key features
Key features of zafu_xiaolin include seamless integration with campus data systems, real-time information retrieval, personalized service offerings based on student data, and secure data processing to protect sensitive information.
Where to use
zafu_xiaolin is primarily used in educational institutions, particularly in campus environments where it can connect with various systems like academic administration, library services, and student information systems.
Clients Supporting MCP
The following are the main client software that supports the Model Context Protocol. Click the link to visit the official website for more information.
Content
农林小林 (XiaoLin) - 基于 MCP 的全栈校园 AI 助手
📝 项目简介
农林小林是面向校园和教育场景的全栈智能助手解决方案框架,专为浙江农林大学开发。灵感来源于论文: HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in HuggingFace。
项目使用自然语言作为大语言模型(LLM)沟通的媒介,集成了任务规划✍️和工具调用🔧为一体,专注于对接现有校园系统,高效解决校园范围内的复杂问题🤔。本项目实现了 Model Context Protocol (MCP) 客户端,使 AI 助手能够无缝连接并利用校园内的各种数据源和工具。
🦾 项目演示
💡 MCP 集成
农林小林实现了 Model Context Protocol (MCP) 客户端,为项目提供了强大的外部数据访问能力。
什么是 MCP?
Model Context Protocol (简称 MCP,中文称为模型上下文协议) 是由 Anthropic 推动的开放标准协议,旨在为大型语言模型 (LLMs) 提供一个标准化接口,使其能够连接并交互外部数据源和工具,克服 LLMs 仅依赖训练数据的局限性。
MCP 核心优势
- 连接外部数据和工具:允许 LLMs 动态获取所需的上下文信息,如数据库记录、文件内容、API 数据等
- 标准化通信协议:定义统一的通信格式,使 AI 应用能够以一致的方式与各种外部系统交互
- 安全性和隐私保护:在本地运行服务器,避免敏感数据上传至第三方
- 灵活性和可扩展性:支持多种数据源和工具的集成,便于构建复杂的 AI 工作流
在农林小林中的应用
作为 MCP 客户端,农林小林能够:
- 校园数据集成:无缝连接学校的各种数据系统,如教务系统、图书馆系统、学生信息系统等
- 实时信息获取:获取最新的校园通知、课程信息、活动安排等
- 个性化服务:根据学生的个人数据提供定制化的学习建议和校园生活指导
- 安全数据处理:确保敏感学生数据在本地处理,保护隐私和安全
MCP 架构集成
农林小林项目采用 MCP 的客户端-服务器架构:
- MCP 客户端:嵌入在农林小林应用中,负责与 MCP 服务器建立通信会话
- MCP 服务器:连接校园各系统,按照 MCP 协议规范暴露标准化功能接口
- 宿主应用:农林小林 AI 助手,集成 MCP 客户端,使模型能够调用校园数据和工具
通过 MCP 集成,农林小林实现了一个真正的"知校园、懂校园"的 AI 助手,能够提供更加精准、实时的校园服务。
💻 项目架构
项目根目录/ ├── backend/ # Django 后端 │ └── chatbot/ # 核心聊天机器人应用 ├── fastapi/ # FastAPI 后端实现 │ ├── app/ │ │ ├── agent/ # 代理系统组件 │ │ ├── api/ # API路由 │ │ ├── core/ # 核心配置 │ │ ├── db/ # 数据库相关 │ │ ├── schemas/ # Pydantic模型 │ │ ├── services/ # 服务层 │ │ └── mcp/ # MCP 客户端实现 │ ├── alembic/ # 数据库迁移 │ └── main.py # 应用入口 └── frontend/ # 前端应用
🚀 运行项目
前提条件
- Python 3.8+
- Node.js 和 npm
- Redis 服务
- MCP 服务器 (用于连接校园数据系统)
1. 运行 Redis
确保 Redis 服务在本地 6379 端口运行:
# Windows用户可使用Docker
docker pull redis
docker run -d -p 6379:6379 --name my_redis redis
2. 启动后端服务
在项目根目录下执行:
# 安装Python依赖
pip install -r requirements.txt
# 启动后端服务
cd backend/
python manage.py runserver
3. 启动前端服务
# 安装前端依赖
cd frontend/
npm install
# 启动前端开发服务器
npm run dev
4. 环境变量配置
修改 backend/chatbot/.env 文件,设置以下变量:
DEEPSEEK_API_KEY=your_deepseek_api_key MCP_SERVER_URL=http://localhost:8080 # MCP 服务器地址 MCP_API_KEY=your_mcp_api_key # 如果需要
或者,如果使用FastAPI版本,创建 fastapi/.env 文件:
DEEPSEEK_API_KEY=your_deepseek_api_key DATABASE_URL=sqlite:///./app.db MCP_SERVER_URL=http://localhost:8080 MCP_API_KEY=your_mcp_api_key
5. 访问应用
- 登录页面: http://localhost:3000/login (默认账号/密码: root/123456)
- 聊天页面: http://localhost:3000/chat
🔄 FastAPI 版本说明
项目提供了基于 FastAPI 的替代后端实现,具有以下特点:
- 异步支持: 原生支持异步编程,适合处理 MCP 请求
- 依赖注入: 使用 FastAPI 的依赖注入系统管理数据库会话和 MCP 客户端
- 类型提示: 使用 Pydantic 模型进行请求和响应验证
- 流式响应: 实现流式聊天响应
- 数据库集成: 使用 SQLAlchemy ORM
- MCP 客户端实现: 集成标准化的 MCP 客户端接口
FastAPI 版本运行
cd fastapi/
pip install -r requirements.txt
# 初始化数据库
alembic revision --autogenerate -m "Initial migration"
alembic upgrade head
# 启动服务
uvicorn app.main:app --reload
API 文档将在以下地址可用:
- Swagger UI: http://localhost:8000/docs
- ReDoc: http://localhost:8000/redoc
💡 未来规划 (RoadMap)
UI界面
- [ ] 手机应用界面
- [ ] 多模态交互
- [ ] 表单输入
后端
- [ ] DFS工作流拓扑
- [ ] RAG系统集成
- [ ] MCP协议扩展支持
- [ ] 多校园系统集成
MCP 相关
- [ ] 校园定制 MCP 服务器开发
- [ ] 多源数据融合与处理
- [ ] 实时数据推送机制
- [ ] 权限与安全控制模块
📞 联系方式
如果您对该项目感兴趣或有任何疑问,可以通过邮箱联系我。
Dev Tools Supporting MCP
The following are the main code editors that support the Model Context Protocol. Click the link to visit the official website for more information.










