window-cap-mcp
A cross-platform MCP server for capturing screenshots of monitors and windows, as well as listing and closing windows.
README
Window Capture MCP Server
A cross-platform window screenshot MCP server and library implemented in Rust, based on rmcp and xcap.
Features
- Get monitor count and details
- Capture screenshots of monitors and windows
- List and close windows
- Cross-platform support (Windows, macOS, Linux)
- Supports multiple transport modes (STDIO, SSE, HTTP)
- Available as Rust library and Python package
Installation
As MCP Server
# Build from source
make build
make run
As Python Package
pip install window-cap-mcp
As Rust Library
[dependencies]
window_cap_mcp_lib = "0.1"
Usage
MCP Server
# STDIO mode (for Claude Desktop)
window-cap-mcp
# SSE mode
window-cap-mcp --sse --port 3000
# HTTP mode
window-cap-mcp --http --port 8080
Python Library
import window_cap_mcp as wc
# Capture monitor screenshot
screenshot = wc.capture_monitor()
# Run as MCP server
wc.run_server()
Rust Library
use window_cap_mcp_lib::{Monitor, Window};
fn main() -> Result<(), Box<dyn std::error::Error>> {
let monitors = Monitor::all()?;
monitors[0].capture_image()?.save("screenshot.png")?;
Ok(())
}
Claude Desktop Configuration
Edit the configuration file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS/Linux:
~/.config/Claude/claude_desktop_config.json
{
"mcpServers": {
"window-cap": {
"command": "window-cap-mcp"
}
}
}
MCP Tools
get_monitor_count
Get monitor information.
{
"count": 2,
"monitors": [
{
"index": 0,
"name": "Display 1",
"width": 1920,
"height": 1080,
"is_primary": true
}
]
}
get_screen_screenshot
Capture monitor screenshot (Base64-encoded PNG).
Parameters: monitor_index (optional)
get_window_list
Get all windows information.
{
"count": 5,
"windows": [
{
"id": 12345,
"title": "VS Code",
"app_name": "Code.exe",
"width": 1280,
"height": 720
}
]
}
get_window_screenshot
Capture window screenshot (Base64-encoded PNG).
Parameters: window_id (required)
close_window
Close a window by ID.
Parameters: window_id (required)
Development
make help # View commands
make build # Build project
make test # Run tests
make clean # Clean artifacts
License
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。