qemu-mcp-server

qemu-mcp-server

An MCP server that gives AI agents direct control over QEMU virtual machines.

Category
访问服务器

README

qemu-mcp-server

npm

An MCP server that gives AI agents direct control over QEMU virtual machines.

Create, boot, snapshot, inspect, and destroy VMs through standard MCP tool calls. The server manages QEMU processes and communicates through QMP (QEMU Machine Protocol).

Quick start

1. Install QEMU

# Ubuntu/Debian
sudo apt install qemu-system-arm qemu-system-x86

# Fedora
sudo dnf install qemu-system-aarch64 qemu-system-x86

# macOS
brew install qemu

# Windows -- use WSL (see "Windows support" below)

2. Install the server

npm install -g qemu-mcp-server

Or run without installing:

npx qemu-mcp-server

Or from source:

git clone https://github.com/Abdalla-Eldoumani/qemu-mcp-server
cd qemu-mcp-server
npm install && npm run build

The server checks for QEMU on startup and tells you exactly what to install if anything is missing.

3. Connect your AI client

Claude Desktop -- add to your MCP config:

{
  "mcpServers": {
    "qemu": {
      "command": "qemu-mcp-server"
    }
  }
}

Cursor -- add to .cursor/mcp.json:

{
  "mcpServers": {
    "qemu": {
      "command": "npx",
      "args": ["qemu-mcp-server"]
    }
  }
}

Claude Code:

claude mcp add qemu -- npx qemu-mcp-server

4. Try it

Ask your AI agent:

"Create an aarch64 VM with 128MB of memory and tell me what you see on the console."

The agent will call create_vm, wait for output, read the console, and report back.

What agents can do

Create and manage VMs

  • create_vm -- Start a VM with specified architecture, memory, and disk/kernel
  • destroy_vm -- Stop and clean up a VM
  • list_vms -- See all running VMs

Control execution

  • pause_vm, resume_vm, reset_vm, shutdown_vm

Snapshots (requires a qcow2 disk image)

  • save_snapshot, load_snapshot, delete_snapshot, list_snapshots

Inspect state

  • get_vm_status, get_vm_info, read_console, dump_memory

Run commands

  • send_console_input -- Type into the serial console
  • wait_for_console_output -- Wait for expected text in output

Example: booting a Linux kernel

If you have a Linux kernel image and root filesystem:

"Create an x86_64 VM with 512MB memory, disk image /path/to/rootfs.qcow2, and kernel /path/to/bzImage with kernel args 'console=ttyS0 root=/dev/vda'."

The agent can then interact with the Linux system through the serial console using send_console_input and wait_for_console_output.

Windows support

This server uses Unix sockets for QMP communication, so it does not run natively on Windows. Use WSL instead:

  1. Install WSL: wsl --install in PowerShell
  2. Install QEMU inside WSL: sudo apt install qemu-system-arm qemu-system-x86
  3. Install Node.js inside WSL
  4. Run the server inside WSL

For Claude Desktop or Cursor on Windows, configure the MCP server to run through WSL:

{
  "mcpServers": {
    "qemu": {
      "command": "wsl",
      "args": ["node", "/path/in/wsl/to/qemu-mcp-server/dist/index.js"]
    }
  }
}

Configuration

All optional. Set via environment variables:

Variable Default What it does
TRANSPORT stdio "stdio" or "http"
HTTP_PORT 3000 Port for HTTP transport
QEMU_BINARY_DIR (empty) Custom path to QEMU binaries
VM_TEMP_DIR os.tmpdir() Where to store sockets and temp files
MAX_VMS 10 Max concurrent VMs
QMP_TIMEOUT_MS 30000 Command timeout (ms)
CONSOLE_BUFFER_LINES 1000 Console output buffer size per VM
LOG_LEVEL info debug, info, warn, error

Supported architectures

  • aarch64 (ARM64)
  • x86_64

Requirements

  • Node.js 18+ (20+ recommended)
  • QEMU installed and on PATH
  • Linux, macOS, or Windows with WSL

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选