Elasticsearch MCP Server

@elastic

代码仓库主页

基本信息

分类

数据库

许可证

Apache-2.0

传输方式

stdio

作者

elastic

关于 Elasticsearch MCP Server

暂无概览

#elasticsearch #mcp #vector-database #mcp-server

Elasticsearch MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "mcp-server-elasticsearch": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "\\"
      ]
    }
  }
}

Elasticsearch MCP Server 工具

List all available Elasticsearch indices

Get field mappings for a specific Elasticsearch index

Perform an Elasticsearch search using query DSL

Execute an ES|QL query

Get shard information for all or specific indices

概览

What is Elasticsearch MCP Server?

The Elasticsearch MCP Server connects AI agents to Elasticsearch data using the Model Context Protocol (MCP), enabling natural language interactions with indices for querying, analysis, and retrieval without custom APIs. This server is deprecated and will only receive critical security updates; it has been superseded by the Elastic Agent Builder MCP endpoint (available in Elastic 9.2.0+ and Elasticsearch Serverless).

How to use Elasticsearch MCP Server?

Deploy the server as a Docker container from AWS Marketplace, then configure it to use either the stdio protocol (direct client communication) or the streamable-HTTP protocol (web-based integrations). Set environment variables (ES_URL, ES_API_KEY or ES_USERNAME/ES_PASSWORD) and run with docker run -i --rm docker.elastic.co/mcp/elasticsearch stdio (or http for HTTP mode).

Key features of Elasticsearch MCP Server

list_indices: List all available Elasticsearch indices
get_mappings: Get field mappings for a specific index
search: Execute a query DSL search
esql: Run an ES|QL query
get_shards: Obtain shard information for indices
Supports stdio and streamable-HTTP transports

Use cases of Elasticsearch MCP Server

Natural language querying of Elasticsearch via AI agents (e.g., Claude Desktop, Cursor, VS Code)
Automated data retrieval and analysis without creating custom APIs
Integrating Elasticsearch data into AI-powered workflows and chatbots
Building search and analytics applications with natural language interfaces

常见问题

Is the Elasticsearch MCP Server still supported?

This server is **deprecated** and will only receive critical security updates. It has been superseded by the [Elastic Agent Builder](https://ela.st/agent-builder-docs) MCP endpoint, available in Elastic 9.2.0+ and Elasticsearch Serverless.

How do I authenticate to Elasticsearch?

Set either `ES_API_KEY` (API key) or both `ES_USERNAME` and `ES_PASSWORD` (basic authentication) as environment variables when running the container.

What protocols does the server support?

The server supports stdio (for direct client connections) and streamable-HTTP (for web-based integrations). Server-Sent Events (SSE) is deprecated—use streamable-HTTP instead.

What are the prerequisites to run the server?

You need an Elasticsearch cluster (8.x or 9.x), Docker installed, an MCP client, and network connectivity between the container and your Elasticsearch cluster. Authentication credentials (API key or username/password) are required.

Where does my data reside?

All data remains in your Elasticsearch cluster; the container does not store data locally. The server communicates with Elasticsearch over HTTPS when using an `https://` URL.