Web Search MCP Server

DuckDuckGo search + web content extraction for AI agents. No API key needed. Free tier works forever.

What agents do with this

• Search the web in real-time for current information
• Fetch and extract readable content from any URL
• Ask questions, verify facts, find sources
• Browse documentation, job listings, news

Quick Install

Method 1: Direct from GitHub (no npm account needed)

Clone and run:

git clone https://github.com/tiagohanna123/web-search-mcp.git
cd web-search-mcp
npm install
npm run build
node dist/index.js

Method 2: npx (if you have npm)

{
  "mcpServers": {
    "web-search": {
      "command": "npx",
      "args": ["-y", "github:tiagohanna123/web-search-mcp"]
    }
  }
}

Method 3: Docker

docker run -i --rm ghcr.io/tiagohanna123/web-search-mcp

Tools

web_search

Search DuckDuckGo. Returns up to 10 results with titles, URLs, and snippets.

Parameters:

• query (required) — The search query
• count (optional, default 5, max 10) — Number of results

web_fetch

Fetch any URL and extract readable content (strips navigation, ads, footers).

Parameters:

• url (required) — The URL to fetch

Premium — R$50 (lifetime)

Get unlimited requests and priority features:

What's included vs Free

Feature	Free	Premium
Requests/minute	10	100
Content extraction	5K chars	50K chars
Cache	5 min	30 min
Rate limit window	1 min	Rolling 24h
Support	GitHub Issues	Priority channel

How to buy

Comprar Acesso Premium — R$50 via Pix ou Cartão

After purchase, email us at [email protected] with your GitHub username to receive a premium access key.

Or self-host with all features (forever free)

The code is MIT licensed. Clone the repo, run it yourself. No limits. Premium is for those who want a managed instance.

Tech Stack

• TypeScript / Node.js
• @modelcontextprotocol/sdk
• duck-duck-scrape (DuckDuckGo API)
• cheerio (HTML parsing)
• Built-in cache (TTL 5 min)