MCP Web Search

MCP server: web search, Wikipedia summaries, and URL content extraction. No API keys required.

Features

• search_web - Two-tier web search (DuckDuckGo HTML / Puppeteer/Bing)
• fetch_url - Extract content with semantic truncation
• summarize_url - Fetch and summarize URL content
• wiki_get - Wikipedia summary by language
• wiki_multi - Wikipedia summaries in multiple languages

Requirements

• Node.js 18+
• Chrome/Chromium (for deep search mode)

MCP Configuration

Claude Code

{
  "mcpServers": {
    "web-search": {
      "command": "npx",
      "args": ["-y", "@zhafron/mcp-web-search"]
    }
  }
}

OpenCode

{
  "mcp": {
    "web-search": {
      "type": "local",
      "command": ["npx", "@zhafron/mcp-web-search"]
    }
  }
}

Generic / Others

{
  "mcpServers": {
    "web-search": {
      "command": "npx",
      "args": ["-y", "@zhafron/mcp-web-search"],
      "env": {}
    }
  }
}

Chrome Setup

| OS | Command | |----|---------| | Ubuntu/Debian | sudo apt install chromium-browser | | Fedora | sudo dnf install chromium | | Arch | sudo pacman -S chromium | | macOS | brew install --cask google-chrome |

Custom path: export CHROME_PATH=/path/to/chrome

Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | CHROME_PATH | auto-detect | Chrome executable path | | HTTP_TIMEOUT | 15000 | Request timeout (ms) | | USER_AGENT | mcp-web-search/1.1 | User agent string |

SSRF Protection: Blocks localhost, 127.0.0.1, ::1, .local domains.

Tools

search_web

Input: { q: string, limit?: number, lang?: string, mode?: "fast"|"deep"|"auto" }

Output: { items: Array<{ title, url, snippet?, source }>, modeUsed, enginesUsed, escalated }

fetch_url

Input: { url: string, mode?: "compact"|"standard"|"full", max_length?: number, format?: "markdown"|"text" }

| Mode | Characters | Tokens | |------|------------|--------| | compact | ~3000 | ~750 | | standard | ~8000 | ~2000 | | full | unlimited | - |

Output: { markdown?, text?, format, url, title?, truncated?, original_length? }

summarize_url

Input: { url: string }

wiki_get

Input: { title: string, lang?: string }

Output: { lang, title, url, description?, extract?, thumbnailUrl? }

wiki_multi

Input: { term: string, baseLang?: string, langs?: string[] }

Troubleshooting

| Issue | Solution | |-------|----------| | Chrome not found | Install Chrome or set CHROME_PATH | | CAPTCHA/blocks | Reduce frequency, use fast mode | | Timeout | Increase HTTP_TIMEOUT |

License

MIT