Security

functional

Server logins, role membership, and permissions in one place.

Server-level security in one dialog: a principal list on the left, an editor form on the right, and a live "Statements to apply" preview underneath. The tabs

follow the engine's own security model rather than forcing one shape onto all of them:
  • SQL Server, aLogins tab (SQL or Windows auth, password with confirm, enable/disable, server-role checklist, and a database-access checklist that creates or drops the matching users) and aUsers tab with its own database picker.
  • PostgreSQL
  • - a single Roles tab: password, can-login, role membership, database connect grants, and schema usage. MySQL / MariaDB - an
  • Accounts tab ('user'@'host', lock state, global and per-database privileges) plus aRoles tab where the server supports roles. ClickHouse - aread-only
  • view over system.users and system.grants that scripts CREATE USER and GRANT statements for you to run. Elsewhere the menu item says why the editor does not apply. SQLite, libSQL / Turso and DuckDB have no logins, users or roles (access is the file's own permissions); Redis manages users with ACL SETUSER / ACL LIST in a query tab; and the wire-compatible labels whose catalogs differ from plain PostgreSQL or MySQL (CockroachDB, YugabyteDB, TiDB, PlanetScale) are left alone rather than guessed at.
Nothing runs until you choose Apply. Every edit is diffed against the loaded baseline into an exact batch of statements (CREATE LOGIN, role grants, per-database CREATE USER), shown verbatim before the Apply (N) button, which counts them. Connections marked production read-only disable Apply entirely.
SQLly
SQLly server security dialog
Server Security: logins and server-role membership, with the permissions each principal holds.

Security also appears on a database node, where it manages that database's users, roles, and grants instead.