home / mcp / google search mcp server

Google Search MCP Server

Provides a read-only MCP server that exposes Google Search data via live queries.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "cdatasoftware-google-search-mcp-server-by-cdata": {
      "command": "java",
      "args": [
        "-jar",
        "/PATH/TO/CDataMCP-jar-with-dependencies.jar",
        "/PATH/TO/Salesforce.prp"
      ]
    }
  }
}

This MCP server provides a local, read-only interface that lets you query live Google Search data exposed through the CData JDBC Driver. You can ask natural-language questions and retrieve up-to-date results without writing SQL, making it easy to empower your AI workflows with Google Search information.

How to use

You connect your MCP client to the local server that runs on your machine. Start the server with the embedded Java runtime, and then ask your AI client to use the built-in tools to read data from Google Search. Typical usage patterns include asking questions like what are the latest results for a given query, or requesting tables and columns to understand the available data. The server handles the data access under the hood, so you don’t need to craft SQL unless you want to for advanced queries.

How to install

Prerequisites: ensure you have Java and Maven installed on your machine. You will build the MCP server, obtain the CData JDBC Driver for Google Search, configure a JDBC connection, and then run the MCP server using a local .prp file.

Step-by-step commands and steps are shown in the order you need to perform them.

Additional steps and configurations

1) Build the MCP server jar to enable the MCP interface.

mvn clean install

# This creates CDataMCP-jar-with-dependencies.jar

2) Obtain and license the CData JDBC Driver for Google Search from the vendor. You will use the driver to connect to Google Search data.

3) Configure your JDBC connection to Google Search. Run the JDBC connection utility to create a connection string, test the connection, and copy the resulting string for later use.

4) Create a properties file that points to your JDBC configuration. Include fields such as Prefix, ServerName, ServerVersion, DriverPath, DriverClass, and JdbcUrl. Leave Tables blank to access all data, or specify the tables you want exposed.

Running the server and using Claude Desktop

To run the MCP server on your machine, execute the Java command and provide the path to your .prp file.

After starting, you can refresh your Claude Desktop client (or any compatible MCP client) to see the new Google Search MCP server, and then begin making data queries through the standard tools.

Notes on usage with Claude Desktop

The server uses a local stdio-based configuration, so it runs on the same machine as the client. If you encounter visibility issues, fully quit and restart the client to refresh the MCP server list.

Available tools

googlesearch_get_tables

Retrieves a list of tables available in the Google Search data source. The output is returned in CSV format with column headers on the first line.

googlesearch_get_columns

Retrieves a list of columns for a specific table. The output is returned in CSV format with column headers on the first line.

googlesearch_run_query

Executes a read-only SQL SELECT query against the Google Search data exposed by the MCP server.