Local-first SQL client for mixed-engine workflows and safe AI help
LakeDB unifies MySQL, MariaDB, SQLite and PostgreSQL access inside a single local desktop workspace, letting you switch engines without separate clients. David Lago builds the app as a utilities tool focused on developer and DBA workflows. Key capabilities include an engine-aware workspace, isolated environment workspaces, and a review-first AI called QuerIA for schema-aware SQL suggestions. The target audience is developers and data engineers who need multi-engine access while keeping credentials and results on their machine.
What the tool does for multi-engine database work
LakeDB opens connections to different engines in one workspace, presenting engine-specific capabilities rather than a generic surface. The app supports MySQL, MariaDB, SQLite and PostgreSQL, so you can view schemas, run queries, and manage multiple environments (development, staging, production) from the same window. The workspace isolates environments so you can keep separate connection sets and switch context without rebuilding settings.
How it handles privacy and sensitive data
The app keeps credentials and query results local by design: LakeDB follows a local-first model where database credentials, row data, and results remain on the user’s device. The documentation states that only schema context required for AI-assisted SQL is sent to the AI service when QuerIA is enabled, and the developer documents that sensitive items otherwise do not leave the machine.
How the AI assistant integrates with user control
QuerIA uses a review-first workflow, meaning AI-generated SQL appears for manual inspection and never executes automatically. The assistant is schema-aware, which lets it generate or correct queries using the database catalog; users must explicitly review and run any suggestion, so execution remains under user control.
Safety, platform behaviour, and installation notes
Production-aware safety features reduce accidental changes: the app enforces read-only modes and environment-specific confirmations to limit risky operations. On Windows, the x64 installer or portable executable may trigger a SmartScreen prompt because beta packages are unsigned; the developer maintains releases on GitHub and offers a winget option for installation.
Who should adopt LakeDB and a practical tip
LakeDB suits developers and DBAs who juggle multiple local or remote engines and need to keep data and credentials local. Use isolated workspaces for each environment and enable read-only enforcement when inspecting production schemas. QuerIA helps write queries but requires manual review before execution, so it fits teams that want AI help without automated changes. Recommended.
