A Model Context Protocol (MCP) server implementation that provides Elasticsearch interaction. This server enables searching documents and analyzing indices.

• list_indices: List all indices.
• get_index: Returns information (mappings, settings, aliases) about one or more indices.
• search_documents: Search for documents.
• get_document: Get a document by ID.

Copy the .env.example file to .env and update the values accordingly.

Start the Elasticsearch cluster using Docker Compose:

docker-compose -f docker-compose-elasticsearch.yml up -d

The default Elasticsearch username is elastic and password is test123.

You can access Kibana from http://localhost:5601.

Add the following configuration to Claude Desktop's config file claude_desktop_config.json. On macOS: ~/Library/Application Support/Claude/claude_desktop_config.json.

{
  "mcpServers": {
    "elasticsearch-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "path/to/src/elasticsearch_mcp_server_project",
        "run",
        "elasticsearch-mcp-server"
      ],
      "env": {
        "ELASTICSEARCH_HOST": "https://localhost:9200",
        "ELASTICSEARCH_USERNAME": "elastic",
        "ELASTICSEARCH_PASSWORD": "test123"
      }
    }
  }
}

Run the MCP server:

uv run mcp_client/client.py src/server.py

Restart Claude Desktop to load the new MCP server.

Now you can interact with your Elasticsearch cluster through Claude using natural language commands like: