Enables web searching through DuckDuckGo's HTML interface to retrieve search results, titles, and snippets.
Provides tools to retrieve page summaries and cross-reference information across multiple languages from Wikipedia.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-web-calcSearch for 'OpenAI Sora' and summarize the most recent news article"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
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
OpenCode
Generic / Others
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