VPN.FAIL MCP
Free remote MCP server for fetching public web pages through a rotating proxy pool. Provides tools to scrape pages as Markdown, text, or HTML, check usage quotas, and service status.
README
<!-- mcp-name: io.github.vpnfail/vpnfail-mcp -->
VPN.FAIL MCP
Free remote MCP server for fetching public web pages through a rotating proxy pool.
Endpoint: https://mcp.vpn.fail/mcp
Access: no account or API key required.
Connect
Use the remote endpoint in an MCP client that supports remote HTTP transport:
{
"mcpServers": {
"vpnfail": {
"type": "http",
"url": "https://mcp.vpn.fail/mcp"
}
}
}
Client configuration formats differ. See client setup and the examples directory before adding it to a specific client.
Available tools
| Tool | Purpose |
|---|---|
scrape |
Fetch one public HTTP(S) page as readable Markdown, plain text, or HTML. |
usage |
View the current anonymous quota state for the requesting IP address. |
service_status |
View public service and proxy-pool availability metadata. |
See exact inputs, result shapes, and errors in docs/tools.md.
Example prompts
Fetch https://example.com and return its readable Markdown.
Fetch this article as plain text and summarize its main claims.
Check the current VPN.FAIL MCP service status.
Public-beta quotas
Anonymous access is limited per source IP address:
| Window | Limit |
|---|---|
| Hour | 250 requests |
| Day | 1,000 requests |
| Week | 5,000 requests |
| Rolling 30 days | 10,000 requests |
There is also a two-request concurrent limit and a 2 MiB maximum response size. See quota details.
Compatibility
The remote endpoint has been tested with the official MCP Inspector CLI. Other client configuration files in this repository are unverified templates, not compatibility claims.
Limitations and responsible use
This is a public beta for fetching public pages. It does not promise JavaScript rendering, CAPTCHA solving, a particular exit location, uninterrupted availability, or access-control bypass. Review limitations and responsible use.
Privacy
No account is required, but this is not a guarantee of anonymity. Requested URLs and request metadata are processed to operate the service. See PRIVACY.md for confirmed behavior and open policy confirmations.
Help
Project scope
This repository publishes public documentation, client configuration examples, and MCP metadata. The hosted service backend is not included here and is not represented as open source.
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。