Web Explorer MCP
Enables private web search and webpage content extraction using a local SearxNG instance, prioritizing user privacy and autonomy.
README
Web Explorer MCP
A Model Context Protocol (MCP) server that provides web search and webpage content extraction using a local SearxNG instance.
Why Web Explorer MCP?
Unlike commercial solutions (GitHub Copilot, Cursor IDE), Web Explorer MCP prioritizes privacy and autonomy:
| Feature | Web Explorer MCP | GitHub Copilot | Cursor IDE |
|---|---|---|---|
| Privacy | ✅ Local SearxNG, zero tracking | ❌ Bing API, Microsoft servers | ❌ Cloud search, third-party APIs |
| Cost | ✅ Free, no limits | 💰 $10-20/month subscription | 💰 $20/month Pro plan |
| API Keys | ✅ None required | ⚠️ GitHub account required | ⚠️ Account & subscription |
| Data Control | ✅ All data stays local | ❌ Queries sent to Microsoft | ❌ Queries sent to external services |
| Setup | ✅ 2 commands | ⚠️ Account setup, policy config | ⚠️ Account, payment setup |
| Open Source | ✅ Fully auditable | ⚠️ Partial (client only) | ❌ Proprietary |
Perfect for: Developers who value privacy, work with sensitive data, or prefer not to depend on external services and subscriptions.
⚠️ Responsible Use
This tool is designed for human-assisted AI interactions, not for automated high-volume scraping:
- 🚫 Not for DDoS - Do not use for overwhelming websites or search engines
- 🚫 Not for High-Speed Automation - Avoid usage speeds significantly higher than a real user
- 🚫 Not for Fully Automated AI Agents - Not recommended for high-performance autonomous agents
- ✅ Respect Infrastructure - Honor website owners' business scenarios and infrastructure capabilities
- ✅ Follow robots.txt - Respect crawling policies and rate limits
Use responsibly: This tool is meant for legitimate research and development, not for abuse.
Features
- 🔍 Web Search - Search using local SearxNG (private, no API keys)
- 📄 Content Extraction - Extract clean text from webpages with Playwright rendering
- 🐳 Zero Pollution - Runs in Docker, leaves no traces
- 🚀 Simple Setup - Install in 2 commands
Quick Start
1. Install Services (SearxNG + Playwright)
git clone https://github.com/l0kifs/web-explorer-mcp.git
cd web-explorer-mcp
./install.sh # or ./install.fish for Fish shell
2. Configure Claude Desktop
Add to your Claude config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"web-explorer": {
"command": "uvx",
"args": ["web-explorer-mcp"]
}
}
}
3. Restart Claude
That's it! Ask Claude to search the web.
Tools
web_search_tool(query, page, page_size)- Search the webwebpage_content_tool(url, max_chars, page)- Extract webpage content with pagination support
Configuration & Usage
See docs/CONFIGURATION.md for:
- Other AI clients (Continue.dev, Cline)
- Environment variables
- Troubleshooting
- Management commands
Update
uvx --force web-explorer-mcp # MCP server
docker compose pull && docker compose up -d # SearxNG + Playwright
Uninstall
docker compose down -v
cd .. && rm -rf web-explorer-mcp
Development
uv sync # Install dependencies
docker compose up -d # Start SearxNG + Playwright
uv run web-explorer-mcp # Run locally
See CONTRIBUTING.md for details.
License
MIT - see 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。