SearXNG MCP Server

by Ihor Sokoliuk

MIT

Repository Docs npm

Featured Web & Search

searchweb-searchsearxngprivacyself-hosted

Summary

Private, self-hosted web search for AI assistants via SearXNG, with pagination, filters, and URL-to-markdown reading.

Features

searxng_web_search with pagination, time filters, language and safe-search
searxng_search_suggestions for autocomplete-style query refinement
searxng_instance_info to inspect available engines/categories
web_url_read fetches and converts a page to markdown

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

Open your agent's MCP configuration file.
Merge the snippet below into it, filling in the environment variables with your own values.
Restart the agent — the "SearXNG MCP Server" tools will be available.

{
  "mcpServers": {
    "searxng-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-searxng"
      ],
      "env": {
        "SEARXNG_URL": "https://your-searxng-instance.example.com"
      }
    }
  }
}

Claude Code

Add the server with a single CLI command.

claude mcp add searxng-mcp-server -e SEARXNG_URL=https://your-searxng-instance.example.com -- npx -y mcp-searxng

Claude Desktop

Add the server to claude_desktop_config.json (Settings → Developer → Edit Config).

{
  "mcpServers": {
    "searxng-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-searxng"
      ],
      "env": {
        "SEARXNG_URL": "https://your-searxng-instance.example.com"
      }
    }
  }
}

Description

This server bridges any MCP client to a self-hosted SearXNG meta-search instance, giving agents privacy-respecting web search instead of relying on a vendor search API. It exposes four tools: searxng_web_search for querying with pagination, time-range, language, and safe-search controls; searxng_search_suggestions for autocomplete-style query refinement; searxng_instance_info to introspect which search engines and categories the connected SearXNG instance has enabled; and web_url_read, which fetches a URL and converts it to clean markdown so an agent can read a result page without a browser.

Because SearXNG aggregates dozens of underlying search engines without tracking, this is a good fit for teams or individuals who want search results without sending every query to a single third-party provider. It runs as a small stdio process that talks to a SearXNG deployment you already run, or connects to a Docker image, configured via a single SEARXNG_URL environment variable. It is one of the most established community search MCP servers, with hundreds of commits, OpenSSF badges, and over a thousand GitHub stars, making it a solid default recommendation for anyone standing up private web search for Claude, Cursor, or other MCP clients.