ArchitectureA supervised JVM-class runtime — OLTP on seven engines, OLAP on three. AI-native, MCP-native, observable as plain SQL.Read the architecture
Está viendo la edición Perú. Está viendo la edición Colombia. You're viewing the Pakistan edition. Cambiar a la edición global →Cambiar a la edición global →Switch to the global edition →

The database workbench gains an AI co-pilot that proposes edits as diffs

An agent tool proposes exact edits to the open tab and the user approves a diff before anything changes. A failed statement can be explained, a buffer translates to the engine's dialect, and Redis and Qdrant get views.

A co-pilot in a database workbench is useful exactly to the degree it cannot surprise you. This release adds one that proposes and waits.

The co-pilot

  • Edits as diffs An agent tool proposes exact string replacements to the open SQL, JavaScript or XML tab. It always requires approval, the user sees the diff, and the browser applies the edit only after a yes.
  • Explain a failure A failed statement's message, stack and script location can be sent to the sidebar for explanation, and the failing line is marked in the editor gutter.
  • Translate A button posts the buffer to the translation service and replaces it with native SQL for the connected engine.

New views

  • Redis A command console with autosuggest over a catalogue of some 960 commands.
  • Qdrant Collections with live status, optimiser state, vectors, indexed vectors, points and segments over the gRPC client.
  • Finding things A database search in the application bar across every server, then tables, views, procedures, functions and sequences inside the chosen one; favourite queries per database; per-statement run marks with language detection.

The approval gate on the edit tool is the same gate every agent tool passes. The workbench is simply where it is most visible.