Drizzle Cube is an open-source semantic layer and embedded analytics toolkit for applications built with Drizzle ORM. A developer defines cubes on top of an existing database schema, including measures, dimensions and relationships. Those definitions can then support an analytics API, React dashboards, a visual analysis builder and AI-assisted data exploration. The public sandbox demonstrates dashboards, notebooks, a schema explorer and a data browser rather than requiring a visitor to imagine the interface from documentation alone.
The project is written around TypeScript and provides server adapters for frameworks including Hono, Express, Fastify and Next.js. Its repository describes support for PostgreSQL, MySQL, SQLite and DuckDB. Client components include an analysis builder and configurable dashboards, with visualizations such as bars, lines, pie charts, funnels and data tables. The documentation also covers funnel, flow and retention analysis, making the toolkit relevant to products that need more than a single reporting chart.
A built-in MCP interface lets compatible AI clients discover cube definitions, validate queries, load data and request chart output. Security context is part of the integration model, including tenant filtering, but developers still need to configure authorization, database permissions and query behavior correctly. A sample tenant filter is not a security audit of a deployed application. The project uses the MIT license and is distributed through npm. Teams should test supported adapters, query plans, joins and performance against their own schema before production use. Drizzle Cube is a developer component for adding analytics to an application, not a hosted replacement that removes responsibility for modeling data or operating the underlying database.
