cube-js/cube
📊 Cube Core is open-source semantic layer for AI, BI and embedded analytics
What it solves
Cube Core provides a standalone, open-source semantic layer that prevents the need to re-implement business logic (metrics, dimensions, and joins) across multiple BI tools, custom applications, or AI agents. It decouples the data model from the visualization layer, ensuring consistency across all downstream consumers of the data.
How it works
Cube Core acts as a headless engine that connects to SQL data sources (such as Snowflake, BigQuery, Databricks, Postgres, Presto, and Amazon Athena). Users define their metrics and dimensions in code, and Cube Core exposes these definitions through SQL, REST, and GraphQL APIs. To ensure high performance and low latency, it includes a built-in relational caching engine.
Who it’s for
It is designed for developers building custom BI experiences, embedded analytics, or AI agents that require a governed semantic foundation for accessing data.
Highlights
- Headless Architecture: No built-in UI, allowing developers to build their own analytics experience.
- Universal Connectivity: Works with all SQL data sources and cloud data warehouses.
- Caching Engine: Built-in relational caching for sub-second latency and high concurrency.
- API-First: Exposes data models via SQL, REST, and GraphQL APIs for easy integration with AI agents and other tools.