MCP.so
ログイン

MCP SearXNG Enhanced Server

@OvertliDS

基本情報

カテゴリ

検索

ライセンス

MIT

トランスポート

stdio

作者

OvertliDS

MCP SearXNG Enhanced Server について

Enhanced MCP server for SearXNG: category-aware web-search, web-scraping, and date/time retrieval.

MCP SearXNG Enhanced Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "mcp-searxng-enhanced": {
      "command": "docker",
      "args": [
        "build",
        "-t",
        "overtlids/mcp-searxng-enhanced:latest",
        "."
      ]
    }
  }
}

MCP SearXNG Enhanced Server のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is MCP SearXNG Enhanced Server?

A Model Context Protocol (MCP) server for category-aware web search, website scraping, and date/time tools. Designed for seamless integration with SearXNG and modern MCP clients.

How to use MCP SearXNG Enhanced Server?

Run via Docker (docker run -i --rm --network=host -e SEARXNG_ENGINE_API_BASE_URL="http://<your-searxng>:8080/search" overtlids/mcp-searxng-enhanced:latest) or natively using Python 3.9+ (pip install -r requirements.txt && python mcp_server.py). Configure SearXNG endpoint and other settings via environment variables. The server can be used in stdio mode (default) or HTTP mode (--http flag).

Key features of MCP SearXNG Enhanced Server

  • SearXNG-powered web search with category support (general, images, videos, files, map, social media)
  • Website content scraping with citation metadata and automatic Reddit URL conversion
  • PDF reading support with conversion to Markdown using PyMuPDF4LLM
  • In-memory caching with automatic freshness validation
  • Domain-based rate limiting to prevent service abuse
  • Timezone-aware date/time tool

Use cases of MCP SearXNG Enhanced Server

  • Build AI assistants that perform category-specific web searches (e.g., images, videos, maps)
  • Scrape and extract content from websites with proper citation metadata
  • Read and convert PDF documents into Markdown for downstream processing
  • Get current date/time in a specified timezone without external API calls
  • Rate‑limit and cache web requests to protect resources and reduce latency

よくある質問

1

What are the prerequisites to run this server?

You need a running SearXNG instance (self-hosted or accessible endpoint) and either Docker installed or Python 3.9+ (Python 3.11 recommended). For native runs, install dependencies with `pip install -r requirements.txt`.

2

How do I configure the server?

Configuration is done via environment variables. The most critical one is `SEARXNG_ENGINE_API_BASE_URL`. You can also use an `ods_config.json` file (environment variables take precedence). See the Environment Variables table in the README for all options.

3

What transport modes are supported?

The server supports stdio transport (default, for MCP clients that spawn a subprocess) and HTTP transport (via FastMCP, for clients connecting via HTTP). In HTTP mode, the server listens on `0.0.0.0:8000` by default and accepts requests at `/mcp`. CORS is fully open.

4

How does caching and rate limiting work?

The server includes in-memory caching with configurable max size (default 100 entries), TTL (default 5 minutes), and maximum age (default 30 minutes). Rate limiting is domain‑based, with defaults of 10 requests per minute and a 60‑second timeout.

5

Can I run the server without Docker?

Yes. Clone the repository, create a virtual environment, install dependencies, set environment variables, and run `python mcp_server.py` (or `python mcp_server.py --http` for HTTP mode). Python 3.9 or newer is required.

コメント

「検索」の他のコンテンツ