- Explore MCP Servers
- AgentX
Agentx
What is Agentx
AgentX is an intelligent agent building platform based on large language models (LLM) and a multi-capability platform (MCP). It aims to simplify the creation process of agents, allowing users to create personalized intelligent agents using natural language and tool integration without complex workflows or drag-and-drop operations.
Use cases
Use cases for AgentX include creating chatbots for customer support, developing virtual assistants for scheduling and reminders, and building educational tools that provide interactive learning experiences.
How to use
To use AgentX, clone the repository from GitHub, set up the PostgreSQL database using Docker, and then build and run the backend service using Maven or Gradle. Detailed installation steps are provided in the README.
Key features
Key features of AgentX include agent management (creation and publishing), LLM context management, agent strategies, a marketplace for tools, a community for MCP Server, and high availability components for models.
Where to use
AgentX can be used in various fields such as customer service automation, personal assistants, educational tools, and any application requiring intelligent conversational agents.
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 Agentx
AgentX is an intelligent agent building platform based on large language models (LLM) and a multi-capability platform (MCP). It aims to simplify the creation process of agents, allowing users to create personalized intelligent agents using natural language and tool integration without complex workflows or drag-and-drop operations.
Use cases
Use cases for AgentX include creating chatbots for customer support, developing virtual assistants for scheduling and reminders, and building educational tools that provide interactive learning experiences.
How to use
To use AgentX, clone the repository from GitHub, set up the PostgreSQL database using Docker, and then build and run the backend service using Maven or Gradle. Detailed installation steps are provided in the README.
Key features
Key features of AgentX include agent management (creation and publishing), LLM context management, agent strategies, a marketplace for tools, a community for MCP Server, and high availability components for models.
Where to use
AgentX can be used in various fields such as customer service automation, personal assistants, educational tools, and any application requiring intelligent conversational agents.
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
AgentX - 智能对话系统平台
AgentX 是一个基于大模型 (LLM) 和多能力平台 (MCP) 的智能 Agent 构建平台。它致力于简化 Agent 的创建流程,让用户无需复杂的流程节点或拖拽操作,仅通过自然语言和工具集成即可打造个性化的智能 Agent。
🔗 相关链接
📦 子仓库
- 🛡️ 高可用网关: API-Premium-Gateway - 模型高可用组件
- 🌐 MCP网关: mcp-gateway - MCP服务统一管理
- 🏪 MCP社区: agent-mcp-community - MCP Server 开源社区
📚 学习资源
- 🎥 项目教程: B站视频教程
- 📖 详细教学: 敲鸭社区 - code.xhyovo.cn
- 🎯 项目演示: 在线PPT介绍
⏳ 功能
- [x] Agent 管理(创建/发布)
- [x] LLM 上下文管理(滑动窗口,摘要算法)
- [x] Agent 策略(MCP)
- [x] 大模型服务商
- [x] 用户
- [x] 工具市场
- [x] MCP Server Community
- [ ] MCP Gateway (等待重构)
- [x] 预先设置工具
- [x] Agent 定时任务
- [ ] Agent OpenAPI
- [x] 模型高可用组件
- [ ] RAG
- [ ] 计费
- [ ] Multi Agent
- [ ] Agent 监控
🚀 如何安装启动
🛠️ 环境准备
必需环境
- Docker & Docker Compose: 用于容器化部署(推荐)
- Git: 用于克隆项目和子模块
本地开发环境(可选)
- Node.js & npm: 推荐使用 LTS 版本
- Java Development Kit (JDK): JDK 17 或更高版本
系统支持
- Linux: 完全支持(推荐)
- macOS: 完全支持
- Windows: 完全支持(Windows 10/11 + WSL2 或原生支持)
🐳 一键启动(推荐)
🔥 开发模式
最佳开发体验:代码修改自动重启容器,无需手动操作!
Linux/macOS 用户
# 克隆仓库
git clone https://github.com/lucky-aeon/AgentX.git
cd AgentX
# 一键启动开发模式(包含热更新功能)
./bin/start-dev.sh
Windows 用户
# 克隆仓库
git clone https://github.com/lucky-aeon/AgentX.git
cd AgentX
# 一键启动开发模式(包含热更新功能)
bin\start-dev.bat
🏭 生产模式
Linux/macOS 用户
# 生产环境启动
./bin/start.sh
Windows 用户
# 生产环境启动
bin\start.bat
📋 服务访问地址
启动成功后,您可以通过以下地址访问服务:
- 前端应用: http://localhost:3000
- 后端API: http://localhost:8080
- API网关: http://localhost:8081
- 数据库连接: localhost:5432
🔐 默认登录账号
系统会自动创建以下默认账号:
角色 | 邮箱 | 密码 |
---|---|---|
管理员 | [email protected] | admin123 |
测试用户 | [email protected] | test123 |
⚠️ 安全提示:首次登录后请立即修改默认密码,生产环境请删除测试账号。
🛠️ 开发管理命令
Linux/macOS 用户
# 查看服务状态
docker compose -f docker-compose.dev.yml ps
# 停止所有服务(保留容器)
./bin/stop.sh
# 删除所有容器
docker compose -f docker-compose.dev.yml down
# 查看服务日志
docker compose -f docker-compose.dev.yml logs -f [服务名]
# 重启特定服务
docker compose -f docker-compose.dev.yml restart [服务名]
Windows 用户
# 查看服务状态 docker compose -f docker-compose.dev.yml ps # 停止所有服务(保留容器) bin\stop.bat # 删除所有容器 docker compose -f docker-compose.dev.yml down # 查看服务日志 docker compose -f docker-compose.dev.yml logs -f [服务名] # 重启特定服务 docker compose -f docker-compose.dev.yml restart [服务名]
📝 开发模式说明
开发模式启动后会显示以下信息并询问是否启动文件监听:
🔥 是否立即启动文件监听?(推荐) - 启动后修改代码会自动重启容器 - 可随时按 Ctrl+C 停止监听 启动文件监听? [Y/n] (默认: Y):
- 选择 Y:启动文件监听,修改代码自动生效
- 选择 n:跳过文件监听,需要手动重启服务
💻 本地开发启动(传统方式)
如果您更喜欢传统的本地开发方式:
1. 启动数据库
cd script
chmod +x setup_with_compose.sh
./setup_with_compose.sh
2. 启动后端服务
cd AgentX
./mvnw spring-boot:run
3. 启动前端服务
cd agentx-frontend-plus
npm install --legacy-peer-deps
npm run dev
功能介绍
Contributors
Star History
联系我们
我们致力于构建一个活跃的开发者社区,欢迎各种形式的交流与合作!
📱 私人微信
如有技术问题或商务合作,可添加开发者微信:

👥 微信交流群
加入我们的技术交流群,与更多开发者一起讨论:

📢 微信公众号
关注我们的公众号,获取最新技术动态和产品更新:

如果二维码过期或无法扫描,请通过私人微信联系我。
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.