databricks -by-cdata
This read-only MCP Server allows you to connect to Databricks data from Claude Desktop through CData JDBC Drivers. For full CRUD support, check out the first managed MCP platform: CData Connect AI (https://www.cdata.com/ai/).
claude mcp add --transport stdio cdatasoftware-databricks-mcp-server-by-cdata java -jar CDataMCP-jar-with-dependencies.jar /PATH/TO/databricks.prp \ --env LD_LIBRARY_PATH="Path to any required native libraries (if applicable)" \ --env JAVA_TOOL_OPTIONS="Optional Java options"
How to use
This MCP server exposes Databricks data through a read-only MCP interface by wrapping the CData JDBC Driver for Databricks. Once running, clients (such as Claude Desktop or other MCP-enabled UIs) can query live Databricks data using the built-in tools exposed by the server, without writing SQL. Tools available include getting the list of available tables, listing columns for a table, and running a SELECT query via the databricks_run_query tool. The server operates over standard MCP/RPC mechanisms and relies on a local Java process that reads from a .prp connection configuration and the JDBC driver. To use it, configure Claude Desktop (or your MCP client) with a server entry that points to the Java runtime and passes the generated databricks.prp as an argument to the jar. Once configured, you can ask natural language questions like “What is the total revenue by region from the Account table?” and the client will invoke the appropriate tools under the hood.
How to install
Prerequisites:
- Java Runtime Environment (JRE) or JDK (recommended 11+)
- Maven (for building from source)
- Access to download the CData JDBC Driver for Databricks
Setup steps:
-
Clone the repository and build the MCP server: git clone https://github.com/cdatasoftware/databricks-mcp-server-by-cdata.git cd databricks-mcp-server-by-cdata mvn clean install
This creates the JAR file: CDataMCP-jar-with-dependencies.jar
-
Download and install the CData JDBC Driver for Databricks from: https://www.cdata.com/drivers/databricks/download/jdbc
-
License the JDBC driver (per driver docs):
- Locate the driver in your installation (DriverPath)
- Run: java -jar cdata.jdbc.databricks.jar --license
- Enter your name, email, and license (or TRIAL)
-
Configure your connection and create a .prp file (e.g. databricks.prp) with your JDBC connection details. Example content to place in the prp file: Prefix=databricks ServerName=CDataDatabricks ServerVersion=1.0 DriverPath=PATH\TO\cdata.jdbc.databricks.jar DriverClass=cdata.jdbc.databricks.DatabricksDriver JdbcUrl=jdbc:databricks:InitiateOAuth=GETANDREFRESH; Tables=
-
Run the MCP server using the generated JAR and the .prp file: java -jar /PATH/TO/CDataMCP-jar-with-dependencies.jar /PATH/TO/databricks.prp
Note: The server uses stdio and is intended for local clients on the same machine
-
If integrating with Claude Desktop or another MCP client, add an entry to the client’s config pointing to the above command and jar, e.g.: { "mcpServers": { "cdatasoftware-databricks-mcp-server-by-cdata": { "command": "PATH\TO\java.exe", "args": [ "-jar", "PATH\TO\CDataMCP-jar-with-dependencies.jar", "PATH\TO\databricks.prp" ] } } }
Additional notes
Tips and caveats:
- The MCP server presented here is configured for read-only access by default. If you need write/update/delete capabilities, you would need a full MCP setup with appropriate driver capabilities enabled and license considerations.
- Ensure the JDBC driver license is valid; otherwise, connections may fail at runtime.
- The server runs via stdio, so it is intended for local clients on the same machine. If you move to a different environment, reconfigure the claude_desktop_config.json (or equivalent) accordingly.
- When testing connections, use the driver’s Connection String utility to verify connectivity and copy the resulting JdbcUrl into the .prp file.
- If the MCP server does not appear in Claude Desktop after starting, fully quit and restart the client to refresh the MCP server list.
