🖥️ Keboola JDBC Driver

The Keboola JDBC driver lets standard SQL clients query data in your Keboola project. It connects through the Query Service API, so any JDBC-compatible tool — DBeaver, DataGrip, or Tableau — can read your tables and platform metadata with plain SQL.

The driver auto-discovers your branches and workspaces, exposes virtual _keboola.* tables for platform metadata, adds a KEBOOLA HELP command, and supports full Snowflake metadata via SHOW commands.

Note

This page is about reading data out of Keboola from external SQL/BI clients. If you instead want to push data into the Tableau ecosystem from Keboola, see the Tableau data destination connector.

Installation guides

Pick your client and follow the step-by-step guide:

• DBeaver
• DataGrip (and other JetBrains IDEs)
• Tableau Desktop

Download the driver

Download the latest keboola-jdbc-driver-X.Y.Z.jar from the GitHub Releases page. Save it somewhere stable on your machine — e.g. ~/keboola/keboola-jdbc-driver.jar.

Prerequisites: Storage API token

The driver requires a Keboola Storage API token. See API Tokens for how to create one.

The token has specific requirements for use with the driver:

• It must not be a bucket-scoped token — the driver needs to list branches and workspaces during connection setup, and a bucket-scoped token cannot see them.
• It needs read access to the buckets you want to query (selecting all buckets is the simplest option).
• Access to components is not required for SQL access.
• The token must come from the same project as the workspace you want to query.

A workspace must also already exist in the project — the driver auto-selects the newest one. Create one in the Keboola UI under Workspaces.

Tip

If you get connection errors, first verify your token works by visiting https://connection.keboola.com/v2/storage/tokens/verify with the header X-StorageApi-Token: <your-token>.

Querying your data

Once connected, you can run standard SQL against your project’s Snowflake tables. The driver also exposes Keboola-specific helpers:

• KEBOOLA HELP lists every Keboola-specific command.
• Virtual _keboola.* tables expose platform metadata. There are five of them: _keboola.components, _keboola.events, _keboola.jobs, _keboola.tables, and _keboola.buckets.

A quick smoke test that confirms the driver is wired up correctly:

KEBOOLA HELP;

SELECT * FROM _keboola.buckets LIMIT 10;

Connection properties

The driver accepts the following connection properties (how you set them depends on the client — see each guide):

Property	Required	Description
token	yes	Your Keboola Storage API token
branch	no	Specific branch ID. Auto-detected (default branch) if omitted
workspace	no	Specific workspace ID. Newest workspace is auto-selected if omitted

The JDBC URL takes the form jdbc:keboola://<host>, where <host> is your Keboola stack — e.g. jdbc:keboola://connection.keboola.com or jdbc:keboola://connection.eu-central-1.keboola.com.

Need Help?

For further help, reach out via Keboola Support.