rasterly-mcp
rasterly-mcp
为任何AI智能体赋予观察和阅读实时网页的能力。这是一个MCP服务器,让您的智能体能够将URL转换为干净的Markdown、可查看的截图或PDF文件——由 rasterly(一个真正的无头Chromium渲染API)提供支持。
适用于任何MCP客户端:Claude Desktop、Claude Code、Cursor、Windsurf等。
工具
工具 | 智能体获得的能力 |
| 将页面的主要内容转换为干净、适合LLM的Markdown格式——JS已渲染,导航/广告/样板内容已剥离。用于阅读URL。 |
| 返回截图图像,智能体可以看到页面(布局、图表等任何视觉元素)。可选 |
| 生成页面的PDF文件(base64资源),适合打印。 |
为什么智能体需要这个:普通的fetch()只能获取原始HTML,而现代JS渲染的网站甚至不会填充内容。read_url会运行一个真正的浏览器,只返回文章内容——令牌更少,信号更强。
Related MCP server: pagewatch
安装
在 rasterly.dev 获取密钥(免费版:每月100次渲染),然后将服务器添加到您的MCP客户端。
Claude Desktop / Claude Code / Cursor — 添加到您的MCP配置文件中
(claude_desktop_config.json、.mcp.json或客户端的MCP设置):
{
"mcpServers": {
"rasterly": {
"command": "npx",
"args": ["-y", "rasterly-mcp"],
"env": { "RASTERLY_API_KEY": "sk_live_your_key" }
}
}
}重启客户端。您的智能体现在可以使用read_url、capture_url和url_to_pdf了。
试试看
向您的智能体提问,例如:
"阅读 https://news.ycombinator.com 并总结前3篇文章。"
"截取stripe.com的截图,告诉我首屏部分说了什么。"
"把这个文档页面转换成PDF。"
配置
环境变量 | 默认值 | 说明 |
| – | 您的rasterly密钥(作为 |
|
| 覆盖以指向自托管的rasterly。 |
独立运行
RASTERLY_API_KEY=sk_live_… npx -y rasterly-mcpMIT · 基于 rasterly 构建
This server cannot be deployed
Maintenance
Related MCP Connectors
Read a URL as clean markdown, screenshot a website, url to PDF. Web access for agents, no signup.
Hosted browser for agents: open pages, search the web, extract data, read PDFs and take screenshots.
URL to screenshot, PDF, or extract for AI agents. Free 150 renders/mo, then $17/mo.
Web tools for AI agents: URL to clean Markdown, site crawl, screenshots, PDF/Office to text, OCR.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables AI agents to fetch any web page as clean markdown or screenshot it, turning URLs into LLM-ready context.21 npmMIT
- AlicenseNot gradedqualityDmaintenanceGives an agent a real browser. Turns any url into clean markdown, takes a screenshot of a url, and converts a url or raw html to pdf. It runs javascript, so it works on pages that a plain fetch returns empty. Respects robots.txt and refuses sites that block automation instead of trying to defeat them. No signup and no api key to start: the first call mints a free trial key and hands it back.MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI agents to crawl and scrape websites, converting HTML to clean Markdown and structured metadata with support for JavaScript rendering, bot evasion, and SSRF protection.192MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to search the web, read and extract content from webpages, fetch JSON from REST APIs, and collect links while bypassing anti-bot protections.22 npmISC