Elasticsearch MCP Server by CData

What is Elasticsearch MCP Server by CData?

Elasticsearch MCP Server by CData is a read-only Model Context Protocol (MCP) server that allows LLMs such as Claude Desktop to query live Elasticsearch data using natural language. It wraps the CData JDBC Driver for Elasticsearch, which exposes Elasticsearch as relational SQL models, and makes that data available through a simple MCP interface — no SQL required.

How to use Elasticsearch MCP Server by CData?

Clone the repository, build the JAR with Maven (mvn clean install), download and license the CData JDBC Driver for Elasticsearch, configure a connection string using the driver's utility, and create a .prp file with JDBC properties. Then add the server entry to your Claude Desktop config file (claude_desktop_config.json) pointing to Java and the JAR, and restart Claude Desktop. The server runs via stdio.

Key features of Elasticsearch MCP Server by CData

• Read-only access to live Elasticsearch data
• Exposes Elasticsearch as relational SQL models via the JDBC Driver
• Supports natural language queries – no SQL needed
• Built-in tools: get_tables, get_columns, and run_query
• Works with LLM clients like Claude Desktop
• Connection configured via a properties (.prp) file

Use cases of Elasticsearch MCP Server by CData

• Ask natural language questions about Elasticsearch data (e.g., "What is the correlation between closed won opportunities and account industry?")
• Query open tickets in a project (e.g., "How many open tickets do I have in the SUPPORT project?")
• Retrieve personal calendar events (e.g., "What calendar events do I have today?")

FAQ from Elasticsearch MCP Server by CData

Is this MCP server read-only?

Yes, this project builds a read-only MCP server. For full read, write, update, delete, and action capabilities, CData offers a separate free CData MCP Server for Elasticsearch (beta).

What are the runtime requirements?

You need Java (to run the JAR file), the CData JDBC Driver for Elasticsearch (installed and licensed), and a client that supports MCP over stdio (like Claude Desktop).

How does the server connect to Elasticsearch?

It uses a JDBC connection string configured in a .prp file. The driver models Elasticsearch data as relational tables and exposes them via MCP tools.

What transport does the server use?

The server uses stdio transport, so it can only be used with clients that run on the same machine as the server.

Where does the data live?

Data remains in the live Elasticsearch instance; the server retrieves it on demand. No data is stored by the MCP server itself.