What are you building?

One SQL interface across connection-backed tables, datasets, and SaaS sources registered in a workspace.

Registering uploaded or query-backed data as tables that appear alongside connection tables in the same query path.

Vector indexes, semantic CLI search with an embedding model, and nearest-neighbor SQL on indexed columns.

Sandboxed sessions for exploratory SQL and datasets separate from shared warehouse or staging environments.

GeoParquet as an uploaded dataset plus spatial predicates (distance, containment) in the same query engine as other tables.

Async query runs, reusing stored results, connection refresh jobs, and listing in-flight work from the CLI.

Using query run history, inspecting tables and columns, and creating sorted or other index types to match access patterns.

Storing data model and team context documents in the workspace for use with agents and the query catalog.

A workspace data model document plus an optional team context document, published and read from the context API.

BM25 full-text indexes, CLI keyword search, and BM25-style SQL alongside relational filters.

Listing connections and tables with columns and types, refreshing after upstream schema changes, and paging wide catalogs from the CLI.