Elasticsearch MCP Server

Add to your client

Copy the config for your MCP client and paste it into its config 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": "<your-elasticsearch-cluster-url>",
        "ES_API_KEY": "<your-elasticsearch-api-key>"
      }
    }
  }
}

Before you start

• An Elasticsearch cluster, version 8.x or 9.x (self-managed, Elastic Cloud, or serverless)
• An Elasticsearch endpoint URL set as ES_URL
• An API key (create one in Kibana under Stack Management > Security > API keys, or via the _security/api_key API) set as ES_API_KEY — or ES_USERNAME / ES_PASSWORD basic-auth credentials
• Node.js (the npm package @elastic/mcp-server-elasticsearch) or Docker installed and running to launch the container

About Elasticsearch MCP Server

Tools & capabilities (5)

When to use it

• Use it when you want to ask questions about data in your Elasticsearch indices in natural language instead of hand-writing Query DSL.
• Use it when you need to discover what indices exist and inspect their mappings before building a query.
• Use it when you want to prototype ES|QL queries interactively from an MCP client.
• Use it when debugging cluster health and you want a quick look at shard distribution.
• Use it when integrating Elasticsearch search results into an agent workflow on an existing 8.x/9.x cluster.

Quick setup

1. 1Confirm you have an Elasticsearch 8.x or 9.x cluster and its endpoint URL.
2. 2Create an API key in Kibana (Security > API keys) or gather basic-auth credentials.
3. 3Add the server to your MCP client, running it via npx (@elastic/mcp-server-elasticsearch) or the Docker image with the `stdio` command.
4. 4Pass ES_URL and ES_API_KEY (or ES_USERNAME/ES_PASSWORD) as environment variables.
5. 5Restart your MCP client and verify by asking the agent to list indices.
6. 6For new deployments on Elastic 9.2.0+, evaluate the Elastic Agent Builder MCP endpoint instead, as this server is deprecated.

Elasticsearch MCP Server FAQ