MCP Server: Elasticsearch semantic search tool

@jedrazb

リポジトリホームページ

基本情報

カテゴリ

メモリとナレッジ

トランスポート

stdio

作者

jedrazb

MCP Server: Elasticsearch semantic search tool について

MCP server to search up-to-date elasticsearch docs

MCP Server: Elasticsearch semantic search tool 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": {
    "elastic-semantic-search-mcp-server": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "\\"
      ]
    }
  }
}

MCP Server: Elasticsearch semantic search tool 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 Server: Elasticsearch semantic search tool?

This Python MCP server enables semantic search on Search Labs blog posts indexed in Elasticsearch. It uses the ELSER model for semantic text retrieval and integrates with Claude Desktop via MCP protocol. Designed for developers building AI-powered search applications with Elasticsearch.

How to use MCP Server: Elasticsearch semantic search tool?

Set ES_URL and ES_AP_KEY in a .env file. Start the server with make dev and access the MCP Inspector at http://localhost:5173. To add to Claude Desktop, run make install-claude-config, which updates claude_desktop_config.json.

Key features of MCP Server: Elasticsearch semantic search tool

Semantic search over Elasticsearch documents using ELSER model
Crawls Search Labs blog posts with Elastic Open Crawler
MCP Inspector interface for testing and debugging
One‑command integration with Claude Desktop
Configuration via environment variables (URL and API key)

Use cases of MCP Server: Elasticsearch semantic search tool

Search indexed blog posts by meaning, not just keywords
Let Claude Desktop retrieve semantic answers from your Elasticsearch corpus
Prototype semantic search workflows with an MCP‑compatible AI assistant
Combine crawling and indexing for a complete search‑pipeline demo

よくある質問

What dependencies does the server require?

You need a running Elasticsearch cluster (with the ELSER model available), the Elastic Open Crawler (Docker image), and an API key with crawl‑and‑search privileges.

Where does the data live?

All documents are stored in an Elasticsearch index named `search-labs-posts`. The crawler populates it; the MCP server queries it.

What transport or authentication does the server use?

The server connects to Elasticsearch via its REST API using the URL and API key supplied in the `.env` file. No additional MCP transport is configured beyond the standard MCP protocol.

What are the known limits?

The crawler and ELSER model must be fully set up before indexing—if using a fresh cluster, wait for the ELSER model to start. The server only works with the `search-labs-posts` index and assumes the mapping defined in the README.