Screenshot MCP Gateway
A unified MCP server that combines screenshot capture with HTTP/SSE gateway for remote access, enabling full-page and element-specific screenshots via Puppeteer.
README
Screenshot MCP Gateway
A unified MCP (Model Context Protocol) server that combines screenshot capture capabilities with HTTP/SSE gateway functionality for remote access. This project integrates HTTP-ScreenShotMCP with Supergateway to provide a single deployable solution for remote screenshot services.
Features
- ✅ Full-page screenshot capture with Puppeteer
- ✅ Element-specific screenshots using CSS selectors
- ✅ Device viewport presets (mobile, tablet, desktop)
- ✅ Custom viewport configurations
- ✅ Advanced wait conditions (selector, function, timeout, networkidle)
- ✅ HTTP/SSE transport for remote access
- ✅ CORS support for cross-origin requests
- ✅ Headless browser operation (perfect for server deployments)
- ✅ Automatic Chrome installation for headless environments
Quick Start
Installation
npm install
This will automatically:
- Install all dependencies
- Download and configure Chromium for headless operation
- Build the TypeScript source
Running the Server
npm start
The server will start on port 3000 (or the port specified in PORT environment variable).
Endpoints:
- SSE MCP endpoint:
http://localhost:3000/sse - Health check:
http://localhost:3000/health
Configuration
Copy .env.example to .env and customize:
cp .env.example .env
Environment Variables:
PORT: Server port (default: 3000)BROWSER_HEADLESS: Run browser in headless mode (default: true)BROWSER_TIMEOUT: Browser timeout in milliseconds (default: 30000)MAX_CONCURRENT_SCREENSHOTS: Maximum concurrent screenshots (default: 5)
Available MCP Tools
1. capture_screenshot
Capture a full-page screenshot of a webpage with advanced options.
Parameters:
url(required): The URL of the webpage to screenshotviewport(optional): Viewport configurationpreset: Device preset (mobile, tablet, desktop)width: Viewport width in pixelsheight: Viewport height in pixelsdeviceScaleFactor: Device scale factorisMobile: Whether to emulate mobile devicehasTouch: Whether device has touch support
waitFor(optional): Wait condition before taking screenshottype: Type of wait condition (selector, function, timeout, networkidle)value: Value for wait conditiontimeout: Timeout in millisecondsidleTime: Network idle time in milliseconds
standardDelay(optional): Apply standard 2.5s delay after networkidle2 (default: true)delay(optional): Additional delay in millisecondswaitUntil(optional): When to consider navigation complete (default: networkidle2)
2. capture_element
Capture a screenshot of a specific element on a webpage.
Parameters:
url(required): The URL of the webpageselector(required): CSS selector for the element to captureviewport(optional): Viewport configurationstandardDelay(optional): Apply standard 2.5s delay (default: true)
3. list_device_presets
List available device presets with their configurations.
No parameters required.
Usage with MCP Clients
Claude Desktop
Add to your Claude Desktop MCP configuration:
{
"mcpServers": {
"screenshot": {
"command": "npx",
"args": [
"-y",
"supergateway",
"--sse",
"http://localhost:3000/sse"
]
}
}
}
Remote Access
To expose your screenshot server publicly, use ngrok:
# Terminal 1: Start the screenshot server
npm start
# Terminal 2: Expose via ngrok
ngrok http 3000
Then use the ngrok URL in your MCP client configuration:
{
"mcpServers": {
"screenshot": {
"command": "npx",
"args": [
"-y",
"supergateway",
"--sse",
"https://your-ngrok-url.ngrok-free.app/sse"
]
}
}
}
Deployment
Standard Deployment
- Clone this repository
- Install dependencies:
npm install - Build the project:
npm run build - Set environment variables (create
.envfrom.env.example) - Start the server:
npm start
Docker Deployment
Create a Dockerfile:
FROM node:18
# Install dependencies for Puppeteer headless Chrome
RUN apt-get update && apt-get install -y \
chromium \
chromium-sandbox \
fonts-liberation \
libappindicator3-1 \
libasound2 \
libatk-bridge2.0-0 \
libatk1.0-0 \
libcups2 \
libdbus-1-3 \
libgdk-pixbuf2.0-0 \
libnspr4 \
libnss3 \
libx11-xcb1 \
libxcomposite1 \
libxdamage1 \
libxrandr2 \
xdg-utils \
--no-install-recommends \
&& rm -rf /var/lib/apt/lists/*
WORKDIR /app
# Copy package files
COPY package*.json ./
# Install dependencies
RUN npm install
# Copy source code
COPY . .
# Build TypeScript
RUN npm run build
# Set environment variables for headless Chrome
ENV PUPPETEER_SKIP_DOWNLOAD=true
ENV PUPPETEER_EXECUTABLE_PATH=/usr/bin/chromium
ENV BROWSER_HEADLESS=true
EXPOSE 3000
CMD ["npm", "start"]
Build and run:
docker build -t screenshot-mcp-gateway .
docker run -p 3000:3000 screenshot-mcp-gateway
Headless Server Requirements
For headless server deployments (Ubuntu/Debian):
# Install Chrome dependencies
sudo apt-get update
sudo apt-get install -y \
chromium-browser \
chromium-codecs-ffmpeg \
fonts-liberation \
libappindicator3-1 \
libasound2 \
libatk-bridge2.0-0 \
libatk1.0-0 \
libcups2 \
libdbus-1-3 \
libgdk-pixbuf2.0-0 \
libnspr4 \
libnss3 \
libx11-xcb1 \
libxcomposite1 \
libxdamage1 \
libxrandr2 \
xdg-utils
Development
Build
npm run build
Development Mode (with auto-reload)
npm run dev
Clean Build Artifacts
npm run clean
Architecture
This project combines:
- Screenshot MCP Server: Provides screenshot capture functionality using Puppeteer
- Supergateway Pattern: HTTP/SSE transport layer for remote MCP access
- Express.js: Web server framework
- MCP SDK: Official Model Context Protocol SDK
- Puppeteer: Headless Chrome automation
Security Considerations
- The server is exposed on the network when running in HTTP mode
- Consider implementing authentication for production deployments
- Use environment variables for sensitive configuration
- Implement rate limiting for production use
- Be cautious with URL inputs to prevent SSRF attacks
Troubleshooting
Chrome Installation Issues
If Chrome fails to install automatically:
npx puppeteer browsers install chrome
Headless Mode Issues
If screenshots fail in headless mode, try:
- Set
BROWSER_HEADLESS=falsefor debugging - Check Chrome dependencies are installed
- Verify Chrome executable path
Connection Issues
- Ensure the port is not already in use
- Check firewall settings for the specified port
- Verify CORS settings if accessing from web clients
License
MIT
Credits
This project integrates:
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。