At a Glance:
Cube Core is an open-source headless semantic layer that defines metrics, dimensions, joins, and access rules in code and exposes them through SQL, REST, and GraphQL APIs for BI tools, custom applications, and AI agents.
Overview:
Cube Core is an open-source semantic layer that centralizes metric and dimension definitions, data source joins, and access rules in a code-based data model. It provides a consistent abstraction layer between SQL data sources and any downstream consumer. This project is designed for teams that need to define business logic once and reuse it across multiple applications without being tied to a single analytics platform. Cube Core supports cloud data warehouses like Snowflake, Databricks, and BigQuery; query engines like Presto and Amazon Athena; and application databases like Postgres. It includes a built-in relational caching engine to handle sub-second query latency and high API concurrency.
Key Decision Points:
Headless with no UI: Cube Core provides no built-in user interface, which means all consumption happens through its APIs — suitable when you need full control over the analytics presentation layer.
API-based consumption layer: Metrics and dimensions are exposed through SQL, REST, and GraphQL endpoints, so any tool or custom application that can consume these protocols can connect.
Works with existing SQL data sources: Cube Core connects directly to a wide range of SQL data stores including cloud data warehouses, query engines, and application databases rather than requiring data ingestion.
Built-in caching engine: A relational caching engine is included to reduce query latency and support high concurrency on API requests without external caching infrastructure.
Full model compatibility with Cube: Data models defined in Cube Core run unchanged in the commercial Cube platform, and vice versa, providing a migration path between self-managed and managed deployments.
Core Features:
Code-defined data model: Metrics, dimensions, joins, and access rules are defined once in code and reused across all consumption channels.
Multi-protocol API access: The semantic layer exposes data through SQL, REST, and GraphQL APIs for different consumer requirements.
Relational caching engine: Built-in caching provides sub-second query latency and handles high concurrency for API requests.
Multi-source SQL connectivity: Works directly with cloud data warehouses, query engines, and application databases without requiring data movement or ingestion.
Headless architecture: No default UI is included, allowing developers to build custom analytics experiences, embedded applications, or AI agent integrations.
Use Cases:
Custom analytics experiences: Development teams building analytics interfaces where the presentation layer must be fully integrated into an existing product design.
AI agent data access: Providing AI agents with a governed semantic foundation that defines consistent metrics and business logic through standard APIs.
Cross-tool metric consistency: Defining metrics once and consuming them across multiple BI tools, internal dashboards, and custom applications without re-implementing models.
Embedded analytics: Building deeply integrated analytics surfaces into applications where the surrounding platform is custom and API access is preferred over a pre-built UI.
Open-Source Alternative Value:
Cube Core provides an alternative to proprietary semantic layers that are tightly coupled to a single BI platform. By exposing the semantic layer through standard APIs while keeping the core open-source, it allows teams to define business logic once and reuse it across different tools and applications. The project is headless by design, which means adoption does not require adopting a complete analytics platform. Data models remain portable between self-managed Cube Core and the commercial Cube offering, giving users flexibility in how and when they adopt managed infrastructure.