home / mcp / json mcp server

JSON MCP Server

This read-only MCP Server allows you to connect to JSON data from Claude Desktop through CData JDBC Drivers. For full CRUD support, check out our MCP Server for JSON (https://www.cdata.com/drivers/json/download/mcp).

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

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

This MCP server provides a read-only interface to live JSON data via the CData JDBC Driver for JSON. It exposes the data through a simple MCP API so you can query JSON sources without writing SQL, making it easy for AI clients to retrieve up-to-date information directly from your JSON data.

How to use

You will run this MCP server on your machine and connect your AI client to it. The server is designed to be used locally and queried through its MCP interface. Once running, you can ask your AI to retrieve tables, inspect columns, or run reads against the JSON data. Use natural language prompts to request information such as counts, aggregates, or specific records, and the client will translate those requests into MCP tool calls under the hood.

How to install

Prerequisites: you need Java installed on your machine. You also need Maven for building the project and access to the CData JDBC Driver for JSON.

1. Clone the project and navigate into it.

git clone https://github.com/cdatasoftware/json-mcp-server-by-cdata.git
cd json-mcp-server-by-cdata

2. Build the MCP server to produce the runnable JAR.

mvn clean install

3. Install the CData JDBC Driver for JSON from the official download page and license it using the provided jar. Follow the driver installation steps to obtain a valid driver JAR and license.

4. Create a connection string for your JSON data source using the CData JDBC Driver utilities, then generate a .prp file with your configuration.

5. Run the MCP server using the generated .prp file.

java -jar /PATH/TO/CDataMCP-jar-with-dependencies.jar /PATH/TO/json.prp

Additional sections

Configuration notes: The server expects an MCP-ready .prp file that includes at least the JDBC connection information (Prefix, ServerName, ServerVersion, DriverPath, DriverClass, JdbcUrl, Tables). This enables the MCP server to expose your JSON data as read-only tools.

Security and access: Run the MCP server on a trusted machine. Restrict access to the local network or use appropriate authentication in your environment to protect sensitive data.

Usage pattern: After starting the server, connect your AI client and begin asking questions about your data. You can request the available tables, inspect columns, and execute read-only queries supported by the MCP tools.

Troubleshooting tips: Ensure the Java runtime is up-to-date and that the .prp file points to a valid JDBC URL. If the MCP server does not appear in the client, fully quit and restart the client so it reloads MCP servers.

Available tools

json_get_tables

Retrieves a list of tables available in the data source. Output is CSV with headers on the first line.

json_get_columns

Retrieves a list of columns for a specified table. Output is CSV with headers on the first line.

json_run_query

Executes a SQL SELECT query against the MCP data source.