Elasticsearch MCP Server for Claude

Overview

The Elasticsearch MCP Server is Elastic's official Model Context Protocol server. It connects Claude (and other MCP clients) directly to an Elasticsearch cluster, so you can search indices, inspect mappings, run ES|QL, and check shard health in natural language — without writing a custom API layer. It is distributed as the docker.elastic.co/mcp/elasticsearch container image and licensed under Apache-2.0.

Key capabilities

The server exposes five tools, each mapping to a real Elasticsearch operation:

• list_indices — list the indices available on the cluster.
• get_mappings — return the field mappings for a specific index.
• search — run a search using Elasticsearch Query DSL.
• esql — run an ES|QL query for tabular, pipe-style analytics.
• get_shards — report shard allocation and health.

How it compares

Several data-store MCP servers let Claude query a backing store; they differ by data model and query language:

Choose the Elasticsearch server when your data lives in Elasticsearch and you want full-text search, relevance ranking, or ES|QL analytics; the PostgreSQL and MongoDB servers cover their own relational and document workloads.

Installation

Claude Code

claude mcp add elasticsearch -- docker run -i --rm \
  -e ES_URL=<your-cluster-url> \
  -e ES_API_KEY=<your-api-key> \
  docker.elastic.co/mcp/elasticsearch stdio

Claude Desktop

Add the server to your configuration file:

{
  "mcpServers": {
    "elasticsearch-mcp-server": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "ES_URL", "-e", "ES_API_KEY",
        "docker.elastic.co/mcp/elasticsearch", "stdio"
      ],
      "env": {
        "ES_URL": "<elasticsearch-cluster-url>",
        "ES_API_KEY": "<elasticsearch-api-key>"
      }
    }
  }
}

To run it as a remote, streamable-HTTP endpoint instead of stdio, start the image with the http argument and publish port 8080:

docker run --rm -e ES_URL -e ES_API_KEY -p 8080:8080 docker.elastic.co/mcp/elasticsearch http

Configuration

Requirements

• Docker, to pull and run the docker.elastic.co/mcp/elasticsearch image.
• A reachable Elasticsearch cluster URL.
• An API key or username/password with access to the indices you want Claude to query.
• An MCP client (Claude Code or Claude Desktop).

Security

• Authenticate with an API key (ES_API_KEY) or basic credentials; scope it to read-only on only the indices Claude should reach.
• Treat ES_URL and credentials as secrets — keep them in the MCP client config or environment.
• ES_SSL_SKIP_VERIFY disables TLS verification and is intended for development only.
• The search and esql tools execute live read queries, so a broad query can add cluster load.

Source Verification Notes

Verified on 2026-06-17:

• The official repository github.com/elastic/mcp-server-elasticsearch documents the docker.elastic.co/mcp/elasticsearch image, the stdio and streamable-HTTP modes, the ES_URL/ES_API_KEY configuration, and the five tools listed above (Apache-2.0).
• Elastic's Search Labs write-up describes the server's purpose and capabilities.
• Claude Code's MCP documentation describes the connector setup pattern used here.