Aichestra Client MCP Server安装配置 GitHub与使用方法

选型结论

Aichestra Client 属于开发工具类 MCP Server，公开资料标注 0 个工具，通过stdio 本地进程接入，客户端适配情况以官方文档为准，有可以直接复制的安装命令；接入前要先准备访问 Token。

分类

开发工具

传输方式

stdio

认证方式

访问 Token

部署方式

公开资料未说明

安装与配置

安装命令

uv run . --agent http://localhost:8001

公开命令不是通用的启动方式，拼不出可靠配置示例；请在客户端 MCP 配置里按官方文档填写 command 和 args。

上手步骤

1
准备运行环境
stdio 服务在本地起进程，需要先装好 Python，再按安装命令拉取依赖。
2
安装服务
复制下面的安装命令执行；命令里带占位符的位置换成自己的密钥或路径。
3
写入客户端配置
在客户端的 MCP 配置里填启动命令和参数，保存后完全退出客户端再打开，让新配置生效。
4
验证是否可用
连接成功后客户端会列出 0 个工具，随便调一个只读工具确认能返回结果。

接入前确认

不要直接授予不必要的文件、网络或账号权限；先核对安装命令和配置内容。

详细介绍

编排器客户端-智能代理命令行界面

用于与Aichestra多代理系统交互的命令行客户端。 最好与智能编排代理 用于无缝代理交互和智能路由。

🚀 智能编排器集成

客户端现在支持 智能代理路由:

自动代理选择

# Intelligent routing to best agent
> "Convert 100 USD to EUR"          → Currency Agent (100% confidence)
> "What is 2+3?"                    → Math Agent (95% confidence)  
> "List ArgoCD applications"        → ArgoCD Agent (100% confidence)

动态代理发现

# Discover available agents
> LIST_AGENTS                       → Show all registered agents
> REGISTER_AGENT:http://localhost:8001 → Register a new agent
> UNREGISTER_AGENT:http://localhost:8001 → Unregister an existing agent
> UNREGISTER_AGENT:currency         → Remove agent

🏗️ 系统集成

客户端与编排器系统集成以实现智能路由：

graph TD
    A[User Input] --> B[🔧 Orchestrator Client]
    B --> C[🤖 Smart Orchestrator]
    C --> D[A2A Card Resolver]
    D --> E[Skill Analysis]
    E --> F{Confidence Scoring}
    F -->|High Score| G[Route to Best Agent]
    F -->|Low Score| H[Fallback Response]
    
    G --> I[☸️ ArgoCD Agent]
    G --> J[💰 Currency Agent]
    G --> K[🧮 Math Agent]
    
    I --> L[Response]
    J --> L
    K --> L
    L --> B
    B --> M[Formatted Output]
    
    subgraph "Intelligent Routing"
        C
        D
        E
        F
        G
    end
    
    subgraph "Agent Ecosystem"
        I
        J
        K
    end

✨ 主要特点

🤖 智能代理路由

自动选择：根据请求内容路由到最佳代理
信心评分：显示路由信心和推理能力
实时发现：动态发现可用代理
基于技能的匹配：将请求与代理功能相匹配

💬 增强的用户体验

自然语言：使用简单的英语进行交互
干净的回应：格式化输出而不是原始JSON
代理发现：显示可用代理及其功能
会话管理：保持对话上下文

🔧 高级客户端功能

WebSocket支持：实时推送通知
历史跟踪：可选对话历史记录
自定义头：支持身份验证和元数据
灵活的配置：多种连接选项

🚀 快速开始

选项1：智能编排器（推荐）

连接到编排器以进行智能代理路由：

# Terminal 1: Start Currency Agent
cd currencyAgent && uv run -m app

# Terminal 2: Start Math Agent  
cd mathAgent && uv run -m app

# Terminal 3: Start ArgoCD Agent
cd argocdAgent && uv run -m app

# Terminal 4: Start Orchestrator
cd orchestrator && uv run -m app

# Terminal 5: Connect Client
cd orchestrator_client && uv run . --agent http://localhost:8000

# Natural language interactions:
# > "Convert 100 USD to EUR"         → Currency Agent (100%)
# > "What is 2+3?"                   → Math Agent (95%)
# > "List ArgoCD applications"       → ArgoCD Agent (100%)

选项2：直接代理连接

直接连接到特定代理：

# Connect to ArgoCD agent directly
uv run . --agent http://localhost:8001

# Connect to Currency agent directly
uv run . --agent http://localhost:8002

# Connect to Math agent directly  
uv run . --agent http://localhost:8003

选项3：定制代理

连接到任何A2A兼容代理：

uv run . --agent http://your-custom-agent:port

📖 使用选项

先决条件

Python 3.10+（A2A SDK兼容性所需）
正在运行的A2A代理服务器（建议使用编排器）
uv包管理器

安装

cd orchestrator_client
uv sync

命令行选项

# Basic usage
uv run . --agent http://localhost:8000

# With custom session ID
uv run . --agent http://localhost:8000 --session 12345

# With history tracking
uv run . --agent http://localhost:8000 --history

# With push notifications
uv run . --agent http://localhost:8000 --use_push_notifications

# With custom headers
uv run . --agent http://localhost:8000 --header "Authorization=Bearer token"

# With custom push notification receiver
uv run . --agent http://localhost:8000 --push_notification_receiver http://localhost:5000

🌐 FastAPI集成

编排器客户端现在支持 FastAPI端点 用于代理管理，提供更快、更可靠的操作。

FastAPI代理管理

# List agents via FastAPI (faster than A2A protocol)
uv run . --agent http://localhost:8000 --list_agent --use_fastapi

# Register agent via FastAPI
uv run . --agent http://localhost:8000 --register_agent http://localhost:8001 --use_fastapi

# Unregister agent via FastAPI
uv run . --agent http://localhost:8000 --unregister_agent "Math Agent" --use_fastapi

# Show API documentation URLs
uv run . --agent http://localhost:8000 --show_api_docs

测试FastAPI集成

# Run comprehensive FastAPI tests
python test_fastapi_integration.py

# Run interactive demo
python test_fastapi_integration.py --interactive

# Show usage help
python test_fastapi_integration.py --help

FastAPI功能

✅ 更快的操作：直接HTTP调用，而不是A2A协议开销
✅ 更好的错误处理：带详细信息的结构化错误响应
✅ 自动回退：如果FastAPI不可用，则返回A2A协议
✅ 交互式文档：Swagger用户界面位于 /management/docs
✅ 混合客户端：支持FastAPI和A2A协议
✅ 类型安全：用于请求/响应验证的Pydantic模型

FastAPI与A2A协议

功能	FastAPI	A2A协议
速度	⚡ 快速	🐌 慢一点
错误处理	📋 结构化	📝 基础
文档	📖 互动	❌ 没有
后退	✅ 是	不适用
类型安全	✅ Pydantic	❌ 有限

🧪 互动体验

代理发现显示

连接后，您将看到可用的代理：

% uv run . --agent http://localhost:8000
Will use headers: {}
======= Agent Card ========
{"capabilities":{"pushNotifications":true,"stateTransitionHistory":false,"streaming":false},"defaultInputModes":["text"],"defaultOutputModes":["text"],"description":"Intelligent agent that routes requests to specialized agents using LangGraph and A2A protocol","name":"Smart Orchestrator Agent","skills":[{"description":"Intelligent request routing to specialized agents","id":"request_routing","name":"Request Routing","tags":["routing","orchestration"]},{"description":"Multi-agent system coordination and management","id":"agent_coordination","name":"Agent Coordination","tags":["coordination","management"]},{"description":"Skill-based agent selection and matching","id":"skill_matching","name":"Skill Matching","tags":["matching","selection"]},{"description":"Confidence scoring for routing decisions","id":"confidence_scoring","name":"Confidence Scoring","tags":["scoring","confidence"]}],"url":"http://localhost:8000/","version":"1.0.0"}

============================================================
🤖 AVAILABLE AGENTS
============================================================
Found 3 available agents:

1. ArgoCD Agent (http://localhost:8001)
   Description: Handles ArgoCD and Kubernetes operations via MCP protocol
   Skills: Kubernetes Management, GitOps, Application Deployment (+3 more)

2. Currency Agent (http://localhost:8002/)
   Description: Handles currency exchange and financial data
   Skills: Currency exchange operations, Financial data analysis, Market analysis and trends (+2 more)

3. Math Agent (http://localhost:8003/)
   Description: Advanced mathematical assistant for calculations, equation solving, calculus, statistics, and matrix operations via MCP
   Skills: Arithmetic Calculation, Equation Solving, Calculus Operations (+2 more)

============================================================
💡 The orchestrator will automatically route your requests to the best agent!
============================================================
=========  starting a new task ========

What do you want to send to the agent? (:q or quit to exit):

智能路由示例

# Financial operations → Currency Agent
> "Convert 100 USD to EUR"
🎯 Routing to Currency Agent (100% confidence)
💰 Response: 100 USD = 85.32 EUR (current rate: 0.8532)

# Mathematical operations → Math Agent  
> "What is 2+3?"
🎯 Routing to Math Agent (95% confidence)
🧮 Response: 2 + 3 = 5

# Kubernetes operations → ArgoCD Agent
> "List all applications"
🎯 Routing to ArgoCD Agent (100% confidence)
☸️ Response: Found 5 applications: guestbook, web-app, api-service...

代理商注册

% uv run . --register_agent http://localhost:8003
Will use headers: {}
======= Agent Card ========
{"capabilities":{"pushNotifications":true,"stateTransitionHistory":false,"streaming":false},"defaultInputModes":["text"],"defaultOutputModes":["text"],"description":"Intelligent agent that routes requests to specialized agents using LangGraph and A2A protocol","name":"Smart Orchestrator Agent","skills":[{"description":"Intelligent request routing to specialized agents","id":"request_routing","name":"Request Routing","tags":["routing","orchestration"]},{"description":"Multi-agent system coordination and management","id":"agent_coordination","name":"Agent Coordination","tags":["coordination","management"]},{"description":"Skill-based agent selection and matching","id":"skill_matching","name":"Skill Matching","tags":["matching","selection"]},{"description":"Confidence scoring for routing decisions","id":"confidence_scoring","name":"Confidence Scoring","tags":["scoring","confidence"]}],"url":"http://localhost:8000/","version":"1.0.0"}
🔄 Registering agent http://localhost:8003 with orchestrator http://localhost:8000
📤 Sending registration request...
🎉 Registration completed successfully!
📄 ✅ Successfully registered Math Agent from http://localhost:8003
Agent ID: Math Agent
Agent Name: Math Agent
Total agents: 3

代理注销

% uv run . --unregister_agent http://localhost:8003
Will use headers: {}
======= Agent Card ========
{"capabilities":{"pushNotifications":true,"stateTransitionHistory":false,"streaming":false},"defaultInputModes":["text"],"defaultOutputModes":["text"],"description":"Intelligent agent that routes requests to specialized agents using LangGraph and A2A protocol","name":"Smart Orchestrator Agent","skills":[{"description":"Intelligent request routing to specialized agents","id":"request_routing","name":"Request Routing","tags":["routing","orchestration"]},{"description":"Multi-agent system coordination and management","id":"agent_coordination","name":"Agent Coordination","tags":["coordination","management"]},{"description":"Skill-based agent selection and matching","id":"skill_matching","name":"Skill Matching","tags":["matching","selection"]},{"description":"Confidence scoring for routing decisions","id":"confidence_scoring","name":"Confidence Scoring","tags":["scoring","confidence"]}],"url":"http://localhost:8000/","version":"1.0.0"}
🔄 Unregistering agent http://localhost:8003 from orchestrator http://localhost:8000
📤 Sending unregistration request...
🎉 Unregistration completed successfully!
📄 ✅ Successfully unregistered Math Agent (ID: Math Agent)
Agent ID: Math Agent
Remaining agents: 2

代理列表

% uv run  . --list_agent
Will use headers: {}
======= Agent Card ========
{"capabilities":{"pushNotifications":true,"stateTransitionHistory":false,"streaming":false},"defaultInputModes":["text"],"defaultOutputModes":["text"],"description":"Intelligent agent that routes requests to specialized agents using LangGraph and A2A protocol","name":"Smart Orchestrator Agent","skills":[{"description":"Intelligent request routing to specialized agents","id":"request_routing","name":"Request Routing","tags":["routing","orchestration"]},{"description":"Multi-agent system coordination and management","id":"agent_coordination","name":"Agent Coordination","tags":["coordination","management"]},{"description":"Skill-based agent selection and matching","id":"skill_matching","name":"Skill Matching","tags":["matching","selection"]},{"description":"Confidence scoring for routing decisions","id":"confidence_scoring","name":"Confidence Scoring","tags":["scoring","confidence"]}],"url":"http://localhost:8000/","version":"1.0.0"}

============================================================
🤖 AVAILABLE AGENTS
============================================================
Found 3 available agents:

1. ArgoCD Agent (http://localhost:8001)
   Description: Handles ArgoCD and Kubernetes operations via MCP protocol
   Skills: Kubernetes Management, GitOps, Application Deployment (+3 more)

2. Currency Agent (http://localhost:8002/)
   Description: Handles currency exchange and financial data
   Skills: Currency exchange operations, Financial data analysis, Market analysis and trends (+2 more)

3. Math Agent (http://localhost:8003/)
   Description: Advanced mathematical assistant for calculations, equation solving, calculus, statistics, and matrix operations via MCP
   Skills: Arithmetic Calculation, Equation Solving, Calculus Operations (+2 more)

============================================================
💡 The orchestrator will automatically route your requests to the best agent!
============================================================

🛠️ 技术实现

A2A协议集成

客户端使用A2A SDK进行标准化的代理通信：

from a2a.client import A2AClient
from a2a.utils import new_agent_text_message

# Create A2A client
client = A2AClient(base_url=agent_url)

# Send message to agent
message = new_agent_text_message(user_input, context_id)
response = await client.send_message(message)

推送通知支持

from orchestrator_client.utils.push_notification_listener import PushNotificationListener

# Start push notification listener
listener = PushNotificationListener(
    host="localhost",
    port=5000,
    notification_receiver_auth=auth
)
listener.start()

会话管理

# Session ID handling
session_id = args.session or str(uuid.uuid4())
context_id = f"session-{session_id}"

# Maintain conversation context
conversation_history = []

📚 发展

项目结构

orchestrator_client/
├── __init__.py
├── __main__.py                    # CLI entry point
├── pyproject.toml                 # Dependencies
├── utils/
│   ├── __init__.py
│   ├── push_notification_auth.py  # Push notification authentication
│   └── push_notification_listener.py # WebSocket listener
└── README.md                      # This file

依赖项

dependencies = [
    "a2a-sdk>=0.2.6,=0.25.0",               # HTTP client
    "starlette>=0.27.0",           # WebSocket server
    "uvicorn>=0.24.0",             # ASGI server
    "pydantic>=2.0.0",             # Data validation
]

🔮 未来的增强功能

GUI界面：用于代理交互的基于Web的用户界面
批处理：支持批处理操作和脚本
代理监控：实时代理健康和性能监控
插件系统：用于自定义功能的可扩展插件架构
多语言支持：支持多种编程语言
高级分析：使用分析和性能洞察

📚 API 参考

主要功能

# Client initialization
client = A2AClient(base_url=agent_url)

# Send message
response = await client.send_message(message)

# Stream responses
async for chunk in client.stream_message(message):
    process_chunk(chunk)

命令行参数

% uv run  . --help
Usage: . [OPTIONS]

Options:
  --agent TEXT
  --list_agent                    List all available agents from orchestrator
  --register_agent TEXT
  --unregister_agent TEXT
  --session INTEGER
  --history BOOLEAN
  --use_push_notifications BOOLEAN
  --push_notification_receiver TEXT
  --header TEXT
  --help                          Show this message and exit.

🤝 贡献

分叉存储库
创建要素分支(git checkout -b feature/amazing-feature)
提交您的更改(git commit -m 'Add amazing feature')
推到分支(git push origin feature/amazing-feature)
打开拉取请求

📄 许可证

该项目是Aichestra多智能体生态系统的一部分。有关详细信息，请参阅主存储库LICENSE文件。

______________________________________________________________________

采用A2A协议、智能路由和无缝代理集成构建 🚀

工具与能力

工具数量

0

资源数量

0

提示词数量

0

能力标签

智能代理Python本地部署命令行工具多代理系统自然语言处理动态路由

数量来自来源目录的标注，具体工具清单以官方文档为准。

典型使用场景

让助手直接读代码

把仓库、提交记录和 issue 交给模型检索，补全和重构时不用来回复制文件。

在对话里跑命令

触发构建、测试和脚本，执行结果直接回到上下文，省去切终端。

接进开发流程

把 PR、CI 状态和代码审查接进助手，减少在几个工具之间跳。

选择建议

值得选的地方

只提供 0 个工具，职责集中，接入后不容易误操作。
有现成的安装命令，不用自己找入口，照着跑就能起服务。
公开说明写得比较完整，接入前能看清能力边界。

先想清楚的地方

stdio 方式要求本地装好 Python，换机器要重新装一遍。
需要自备访问 Token，没有凭据没法直接试。
公开摘要偏短，具体能力和限制要回到仓库确认。

更适合

需要在开发工具场景里让模型直接调用外部能力的人。
想先快速试一下、能接受在本地跑服务的人。

不太适合

完全不想在本地装环境的人。
暂时拿不到访问 Token的人。
需要厂商级 SLA 和技术支持的生产环境。

MCP Server	工具数	传输方式	安装命令	收录来源	最后核验
Aichestra Client MCP World · 本页	0	stdio	有	1 个	2026-05-17
Git 838e2ec1 MCP World	12	stdio	有	2 个	2026-05-17
Markitdown MCP 2c653a5f MCP World	1	stdio	有	1 个	2026-05-17
Context7 MCP C671e75a MCP World	0	stdio	有	1 个	2026-05-17
Context7 831acf40 MCP World	0	stdio	有	1 个	2026-05-17

常见问题

5 条

Aichestra Client 怎么安装？

公开资料给出的安装命令是 uv run . --agent http://localhost:8001，运行环境为 Python。装完后按客户端要求写进 MCP 配置，重启客户端生效。

Aichestra Client 怎么在 Claude Desktop、Cursor 里配置？

它是stdio 本地进程，客户端配置里要写启动命令和参数，运行环境是 Python；保存后完全退出客户端再打开，配置才会生效。

Aichestra Client 收费吗？

本站只整理公开资料，不代收费用，也没有标注具体价格。是否收费取决于服务本身和背后的接口额度，正式使用前按官方说明确认。

Aichestra Client 需要 API Key 吗？

公开资料标注的认证方式是访问 Token，也就是要先拿到凭据才能连上；凭据只填在自己本地的配置里，不要写进公开仓库。

Aichestra Client 支持哪些客户端？

公开资料没有标注具体客户端；能配置本地 MCP 命令的客户端都能接，配置方式参照官方文档。

来源信息

MCP World

aichestra-client

用于与多代理系统交互的命令行客户端，支持智能代理路由、动态代理发现和自然语言交互。

www.mcpworld.com

/detail/814cb09e1e29221d4653589a6f62cc7c

参考链接

仓库入口

github.com

/aichestra/aichestra-client

来源标签