Multi-Adapter Support
One interface, many platforms. Connect LakeQL to the data engines your team already runs on.
Trino
A type-safe GraphQL-over-HTTP layer over Trino. No driver to install, with generated SQL and automatic field mapping.
Microsoft Fabric
Unified type-safe GraphQL over Fabric warehouses and lakehouses, with request-scoped Entra ID identity and storage-agnostic access.
ClickHouse
Coming soonRun sub-second analytical queries against ClickHouse clusters with native protocol batching.
DuckDB
Coming soonQuery local or in-memory DuckDB files and Parquet datasets directly from your app or CLI.
Powerful features
Everything you need to query, govern, and ship data products faster.
Type-safe GraphQL APIs
Adapter-agnostic query generation with full TypeScript support. Get autocomplete and compile-time checks for your entire data layer.
CLI-based schema generation
Pull live catalog metadata from multiple sources and generate typed schemas that stay in sync as upstream tables evolve.
Filtering, sorting & pagination
Built-in support for expressive filters, sorting, and pagination with engine-specific SQL dialect generation.
Adapter-agnostic permissions
Define access rules per data source and table name. Enforce them consistently across every connected adapter.
OpenTelemetry tracing
Optional per-DataSource tracing out of the box. Monitor query latency, errors, and throughput with your existing observability stack.
Flexible authentication
OAuth2, Basic Auth, and Bearer token authentication with configurable impersonation strategies per DataSource.
Free & open source
LakeQL is Apache 2.0 licensed and built in the open. Read the source, file an issue, or ship a new adapter — every contribution makes data access better for everyone.