SQL clients
A desktop SQL client does not read your bucket by itself. It connects to an engine: either DuckDB running inside the client, which reads your bucket or table directly with a short setup script, or a warehouse such as Trino, Athena or BigQuery that you have already pointed at the data. Pick DuckDB when you have no warehouse; it needs only a read-only storage key.
| Tool | Connect through | Page |
|---|---|---|
| DBeaver | DuckDB driver with an init script; Trino, Athena, BigQuery, ClickHouse, PostgreSQL or MySQL connection | DBeaver |
| DataGrip and other JetBrains IDEs | DuckDB data source with a startup script; Trino, Athena, BigQuery, ClickHouse, PostgreSQL or MySQL data source | DataGrip |
| TablePlus | DuckDB connection with bootstrap commands; BigQuery, ClickHouse, PostgreSQL or MySQL connection | TablePlus |
| DuckDB UI and Harlequin | DuckDB started with an init script; Harlequin also has Trino, PostgreSQL and BigQuery adapters | DuckDB UI and Harlequin |
Every DuckDB path runs the same statements: the CREATE SECRET, ATTACH and CREATE VIEW for your destination from DuckDB, saved in a file the client runs when it connects. After that, the example queries work unchanged.