XiaoZhi Windows MCP Agent Bridge
# XiaoZhi Windows MCP Agent Bridge
XiaoZhi Windows MCP Agent Bridge 是将桌面上带有麦克风与扬声器的 ESP32 小智 AI 设备 连接到本地 Windows 电脑的桥接服务。
通过标准的 Model Context Protocol (MCP),小智可以作为语音指挥官 (Voice Commander),调用 Windows 本地工具、执行电脑操作、全盘读写文件、控制 Edge 浏览器、截取屏幕进行视觉分析,并重点调度本地 Codex CLI / Claude Code / OpenCode 等 AI Agent 矩阵执行重型开发与工程任务。
整体架构:语音指挥官 + 本地 Agent 矩阵
小智 ESP32 硬件 (麦克风/喇叭/屏幕)
↓ (语音交互)
xiaozhi.me 官方云端 (ASR / LLM / TTS / MCP Client)
↓ (WebSocket JSON-RPC 2.0)
XiaoZhi Windows Agent Bridge (本机 Python 进程)
↓
Tool Registry (注册 49 个本地能力工具)
↓
┌─────────────────┬─────────────────┬─────────────────┬─────────────────┐
│ Windows 应用 │ 本地文件系统 │ 终端与 PowerShell│ 多 Agent 调遣器 │
│ • 智能启动/搜索 │ • 全盘/目录遍历 │ • Safe Terminal │ • Codex CLI 主力│
│ • 进程聚焦/退出 │ • 文件读写/删除 │ • 任意 Shell │ • Claude / Open │
├─────────────────┼─────────────────┼─────────────────┼─────────────────┤
│ Desktop GUI │ 屏幕视觉 (Vision│ Edge 浏览器 │ Skills/Workflow │
│ • 窗口枚举/点击 │ • 内存屏幕截图 │ • 网页搜索/打开 │ • 声明式 Skill │
│ • 文本输入/按键 │ • MCP Image 回传│ • 页面提取/点击 │ • YAML 工作流 │
└─────────────────┴─────────────────┴─────────────────┴─────────────────┘Related MCP server: WinCommander
核心功能与语音指令
功能类别 | 语音指令示例 | 底层调用工具 |
Agent 调度 | “小智,让 Codex 帮我把 ESP32 项目加上 Wi-Fi 重连逻辑” |
|
Agent 进度 | “小智,刚才 Codex 的任务完成了吗?” |
|
Agent 清单 | “小智,我电脑里有哪些可以干活的 Agent?” |
|
软件启动 | “小智,打开 VS Code” / “打开飞书” |
|
屏幕视觉 | “小智,看看我电脑屏幕上有什么” |
|
文件查找 | “小智,看看桌面上有哪些 PDF” |
|
文件读写 | “小智,帮我读取这个配置 / 写一个贪吃蛇” |
|
项目 Git | “小智,看看我的项目最近改了什么” |
|
终端执行 | “小智,运行这个项目的测试 / 查看 python 版本” |
|
浏览器交互 | “小智,搜索 ESP32-S3 数据手册” |
|
快速配置与启动
1. 环境准备
Windows 10 / 11
Python 3.11+
(可选)已安装的 Codex CLI (
npm i -g @openai/codex)、Claude Code 等
安装依赖:
pip install -r requirements.txt2. 配置 MCP Endpoint
复制 .env.example 为 .env 并填入从小智控制台获取的 Endpoint:
XIAOZHI_MCP_ENDPOINT=wss://api.xiaozhi.me/mcp/?token=YOUR_TOKEN_HERE3. 配置 Workspace(可选)
在 config.yaml 中添加你的本地开发工程路径:
workspaces:
agent:
path: .
esp32:
path: D:/Projects/esp32
stm32:
path: D:/Projects/stm324. 运行 Agent
python main.py终端将输出:
===================================
XiaoZhi Windows Agent
===================================
XiaoZhi MCP CONNECTING
Tools 49
Security ENABLED
Permission UNRESTRICTED
MCP server handshake completed
Connected; registered 49 local tools自动化测试
运行完整测试套件:
python -m pytestThis server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- FlicenseNot gradedqualityCmaintenanceConnects a Xiaozhi AI voice agent to a Windows PC, enabling autonomous control via 58 MCP tools for screen vision, web browsing, file management, and system control.
- FlicenseNot gradedqualityBmaintenanceTurns any Windows device into a remotely controllable MCP toolset, allowing a mobile AI agent to execute CLI, GUI, browser, and system commands on Windows without an API key.2
- FlicenseNot gradedqualityBmaintenanceEnables AI agents to securely control a remote Windows/Linux PC via MCP and REST, executing shell commands and driving coding-agent CLIs like Claude, Cursor, and Codex.
- FlicenseNot gradedqualityCmaintenanceMCP server enabling voice-controlled Windows PC automation via Xiaozhi AI, supporting 200+ apps, media playback control, and terminal command execution.
Related MCP Connectors
Personal assistant MCP server with search, execute, packages, jobs, secrets, and integrations.
Hosted MCP server connecting claude.ai, ChatGPT and other AI apps to your own computer
OCR, transcription, file extraction, and image generation for AI agents via MCP.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Joho6666/xiaozhi-windows-agent'
If you have feedback or need assistance with the MCP directory API, please join our Discord server