▌BestMCPTools.org
Elasticsearch MCP logo

Elasticsearch MCP

Search indices and run queries against Elasticsearch from an agent.

SearchFreeClaudeClaude CodeCursorVS CodeAll ModelsModerate✓ Verified
Visit Tool719 stars

Last verified: September 20, 2026

Own Elasticsearch MCP? Boost it to the top.✦ Upgrade to Featured

Maintenance status

Active
Last code update: September 19, 2026License: Apache-2.0Checked September 28, 2026

How to Install Elasticsearch MCP

$npx -y @elastic/mcp-server-elasticsearch

Requires Claude Desktop, Cursor, Windsurf, or another MCP-compatible client.

Setup by client

Claude Desktop (claude_desktop_config.json)

{
  "mcpServers": {
    "elasticsearch-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@elastic/mcp-server-elasticsearch"
      ]
    }
  }
}

Cursor (~/.cursor/mcp.json)

{
  "mcpServers": {
    "elasticsearch-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@elastic/mcp-server-elasticsearch"
      ]
    }
  }
}

VS Code (.vscode/mcp.json)

{
  "servers": {
    "elasticsearch-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@elastic/mcp-server-elasticsearch"
      ]
    }
  }
}

Claude Code (terminal)

claude mcp add elasticsearch-mcp -- npx -y @elastic/mcp-server-elasticsearch

Add any API keys or environment variables the server's README lists.

About Elasticsearch MCP

Elastic's MCP server lets an agent list indices, inspect mappings and run search queries against an Elasticsearch or OpenSearch cluster in natural language. For log-heavy environments this collapses the usual loop of writing a query DSL blob by hand, and it pairs well with observability workflows where the agent already has error context. Read-only configuration is strongly recommended in production.

Elasticsearch MCP is a Search MCP server designed for researchers, analysts, and engineers building grounded agents. It enables giving AI assistants live, source-grounded web access by wrapping a search provider and page-fetch step as MCP tools. Key capabilities include live web search with snippets and urls, optional full-page fetch for cited sources, configurable result counts and freshness windows, and structured tool calls that work cleanly inside an LLM context. It integrates with Google, Bing, Brave, Tavily, Exa, Perplexity, and SerpAPI, and is best suited for researchers, analysts, and engineers building grounded agents who need research and competitive analysis.

Key Features

  • elasticsearch
  • search
  • logs
  • observability
  • query-dsl

Pricing

Free: from Free (open source; Elasticsearch cluster required)

Pros & Cons

Pros

  • Lets the assistant answer about events past its training cutoff
  • Cites sources you can actually click
  • Replaces most simple RAG pipelines

Cons

  • Quality depends on the underlying search provider
  • Token costs scale with how much page content you pull
  • Rate limits can bite on bursty agent runs

Best For

  • Research and competitive analysis
  • Source-grounded answers in chat UIs
  • Fact-checking long-form drafts
  • Replacing manual Googling in agent loops

Our Take on Elasticsearch MCP

Removes the pain of hand-writing query DSL. Lock it to read-only credentials before pointing it at a production cluster.

Frequently Asked Questions

Tags

#Search#elasticsearch#search#logs#observability#query-dsl