SQL and command editors

Tabs

Each SQL tab owns its query text, connection and schema, result page, loading state, cancellation request and error. Several tabs can run at the same time. Closing one tab cancels its request without stopping unrelated queries.

Completion and highlighting

Syntax and completion follow the selected engine dialect. PostgreSQL, MySQL and SQLite completion includes table names and columns throughout the selected schema, with schema-qualified names and alias-aware suggestions. Identifiers are quoted according to the engine.

Running statements

• In result mode, SELECT and CTE statements run separately and the final statement's result is shown and paginated. Use separate tabs to keep several result sets.
• Use Execute statement for DDL or writes. The UI asks for confirmation and reports affected rows. Statements run separately in order without an implicit transaction; if a later statement fails, the error identifies it and earlier statements remain committed.
• Statement splitting understands quoted strings and identifiers, comments and PostgreSQL dollar-quoted bodies. Client dump directives such as MySQL DELIMITER and PostgreSQL COPY FROM STDIN are not supported.
• Queries have a 30-second timeout; connection tests have a 15-second timeout. Cancellation propagates to the database cursor.

History

The History tab in the sidebar records SQL, connection and duration for the last 50 statements. It is stored locally in the webview and never includes results.

MongoDB and Redis

Document and key engines get native command tabs rather than pretend SQL. MongoDB accepts Extended JSON find and aggregate commands for paginated results, and explicit Execute mode for other commands. Redis accepts shell-style commands or JSON argument arrays. Documents are pretty-printed with syntax highlighting.