Build Databricks-Powered Applications in Claude Code with CData Code Assist MCP

Claude Code is an AI-powered command line tool that enables agentic coding workflows. With support for MCP, Claude Code can connect to local tools and enterprise data sources directly from your terminal, enabling natural language interaction with live systems without switching context.

Model Context Protocol (MCP) is an open standard for connecting LLM clients to external services through structured tool interfaces. MCP servers expose capabilities such as schema discovery and live querying, allowing AI agents to retrieve and reason over real-time data safely and consistently.

The following steps cover installing the CData Code Assist MCP for Databricks, configuring the connection to Databricks, connecting the Code Assist MCP add-on to Claude Code, and querying live Databricks data from within the terminal.

Getting Started

---

Step 1: Download and install the CData Code Assist MCP for Databricks

1. To begin, download the CData Code Assist MCP for Databricks 👁 Downloading the CData Code Assist MCP add-on installer
   
2. Find and double-click the installer to begin the installation
3. Run the installer and follow the prompts to complete the installation 👁 Completing the CData Code Assist MCP add-on installation
   

When the installation is complete, you are ready to configure your Code Assist MCP add-on by connecting to Databricks.

Step 2: Configure the connection to Databricks

1. After installation, open the CData Code Assist MCP for Databricks configuration wizard 👁 The CData Code Assist MCP add-on configuration wizard
   

   NOTE: If the wizard does not open automatically, search for "CData Code Assist MCP for Databricks" in the Windows search bar and open the application.

   👁 Opening the CData Code Assist MCP add-on configuration wizard (Google Sheets is shown).
   
2. In MCP Configuration > Configuration Name, either select an existing configuration or choose to create a new one 👁 Creating a new configuration
   
3. Name the configuration (e.g. "cdata_databricks") and click OK 👁 Naming the configuration
   

4. Enter the appropriate connection properties in the configuration wizard

   To connect to a Databricks cluster, set the properties as described below.

   Note: The needed values can be found in your Databricks instance by navigating to Clusters, and selecting the desired cluster, and selecting the JDBC/ODBC tab under Advanced Options.

   • Server: Set to the Server Hostname of your Databricks cluster.
   • HTTPPath: Set to the HTTP Path of your Databricks cluster.
   • Token: Set to your personal access token (this value can be obtained by navigating to the User Settings page of your Databricks instance and selecting the Access Tokens tab).
   👁 Configuring the Code Assist MCP add-on connection
   
5. Click Connect to authenticate with Databricks
6. Then, click Save Configuration to save the Code Assist MCP add-on 👁 Saving the Code Assist MCP add-on configuration
   

This process creates a .mcp configuration file that Claude Code will reference when launching the Code Assist MCP add-on. Now with your Code Assist MCP add-on configured, you are ready to connect it to Claude Code.

Step 3: Connect the Code Assist MCP add-on to Claude Code

1. Install the Claude Code CLI using the terminal
2. Open the Claude Code configuration file at ~/.config/claude-code/config.json (or the location shown after initialization)

Option 1: Manually add the MCP configuration

1. Open the mcp.json file in your preferred editor
2. Add the code shown below

{
"mcpServers" : {
 "cdata_databricks" : {
 "type" : "stdio",
 "command" : "C:\Program Files\CData\CData Code Assist MCP for Databricks\jre\bin\java.exe",
 "args" : [ "-Dfile.encoding=UTF-8", "-jar", "C:\Program Files\CData\CData Code Assist MCP for Databricks/lib/cdata.mcp.databricks.jar", "cdata_databricks" ], 
 "env" : {}
 }
}
}

NOTE: The command value should point to your Java 17+ java.exe executable, and the JAR path should point to the installed CData Code Assist MCP add-on .jar file. The final argument must match the MCP configuration name you saved in the CData configuration wizard (e.g. "cdata_databricks").

Option 2: Copy the MCP configuration from the CData Code Assist MCP for Databricks UI

1. After saving and testing your connection in the configuration wizard, click Next
👁 Selecting 'Next'

2. Select Claude Code from the AI MCP Tool dropdown
👁 Choosing the required AI MCP tool

3. Click Copy JSON to copy the generated MCP configuration to your clipboard
👁 Copying the MCP configuration

4. Paste the copied JSON into the mcp.json file

Step 4: Verify connection in Claude Code

Claude Code provides tools to verify the connection is active before building.

1. Open a terminal and navigate to your project directory. Run the command claude mcp list
2. Check that your configuration name appears with a Connected status
👁 Checking the connection status

3. Start Claude Code by running claude
4. Inside the Claude Code session, type /mcp to view active servers 👁 Viewing active MCP servers
   

Step 5: Query live Databricks data in Claude Code

With the connection verified, you can now use natural language prompts to query and work with live Databricks data.

1. Prompt Claude Code to review the instructions for your MCP connection to ensure it has all the appropriate context when writing code 👁 Reviewing MCP connection instructions
   
2. Start building with natural language prompts! For example:

   For my project, data from the Customers is very important. Pull data from the most important columns like City and CompanyName.

Claude Code will use the MCP add-on to connect to Databricks, retrieve the requested data, and provide results directly in your terminal

Build with Code Assist MCP. Deploy with CData Drivers.

Download Code Assist MCP for free and give your AI tools schema-aware access to live Databricks data during development. When you're ready to move to production, CData Databricks Drivers deliver the same SQL-based access with enterprise-grade performance, security, and reliability.

Visit the CData Community to share insights, ask questions, and explore what's possible with MCP-powered AI workflows.