広告掲載
MCP.so
ログイン

MCP Server Search

@MaartenSmeets

リポジトリ

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

MaartenSmeets

MCP Server Search について

MCP server to use a search engine to obtain locations of relevant information on the internet

MCP Server Search 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-server-search": {
      "command": "docker",
      "args": [
        "build",
        "-t",
        "mcp-server-search",
        "."
      ]
    }
  }
}

MCP Server Search のツール

1

Search Google and return results. Results are presented in a formatted Markdown structure with titles, URLs, and descriptions (when enabled).

概要

What is MCP Server Search?

MCP Server Search is an MCP (Model Context Protocol) server that provides Google search functionality for AI models. It allows models to search for up-to-date information from the web, with caching, configurable delays, and retry mechanisms to avoid rate limiting.

How to use MCP Server Search?

Install via Docker: build the image, create ./logs and ./cache directories, and run the container with SSE enabled on port 8000. Configure Cline to connect to http://localhost:8000/sse. Use the google_search tool or prompt with a query and optional parameters like num_results (1–20), use_cache, and include_descriptions.

Key features of MCP Server Search

  • Google search integration with file-based caching
  • Configurable request delays and retries
  • Support for simple and advanced searches (with descriptions)
  • Throttling and randomization to be a good web citizen
  • Exponential backoff on rate limiting (HTTP 429)
  • Thread-safe access to cached results

Use cases of MCP Server Search

  • AI models retrieving real-time web information
  • Searching for current events or data without manual browsing
  • Integrating search results into conversational AI responses
  • Caching frequent searches to improve performance

よくある質問

1

What dependencies does the server require?

Docker is the recommended installation method; the server runs on Python with dependencies defined in the Dockerfile.

2

Where are search results cached?

Results are cached in a file-based database at `cache/google_cache.db` (configurable via `--cache-path`), with thread-safe access.

3

What transport does the server use?

The server uses HTTP SSE (Server-Sent Events) on port 8000 at the `/sse` endpoint.

4

Can I configure the search delay and retries?

Yes, via command-line arguments: `--request-delay` (default 5 seconds) and `--max-retries` (default 3).

5

Is there rate limiting protection?

Yes, the server implements automatic request throttling, random user agent rotation, and exponential backoff on HTTP 429 responses.

コメント

「検索」の他のコンテンツ

Fitralo

Fitralo

Search Europe's road freight carriers from your AI assistant: find carriers by route, equipment and certificates among 15,700+ checked carriers in 32 countries, in 14 languages. Read-only, no account needed for searching

Google Search MCP Server

HasData

Hosted Google Search / SERP MCP server from HasData, no Google Cloud project. Organic results, AI Overview, People Also Ask, AI Mode, news and shopping for Claude, Cursor and any MCP client.

23

Webz.io News Search

Webz.io

Search global news using natural language. Webz.io News Search API returns the most relevant articles and content, with filters for source, country, language, date, sentiment, and category.

Default Privacy

Default Privacy

Sourced privacy answers for AI assistants: how private an app or service is, private replacements for mainstream apps, privacy tool facts with sources, guides, MCP server vetting and free domain checks. Read-only, no acc

3

3GPP Scout

3gppscout

Hosted 3GPP MCP server: cited search across Rel-15 through Rel-20 TS/TR specs, recent TDocs, company proposals and meeting decisions. OAuth 2.1; no local index required.

広告掲載