Cube.js is now called Cube Core: an open-source semantic layer that connects data sources to BI tools, custom applications, and AI agents. It centralizes metric definitions, dimensions, joins, and access rules, then serves that governed data through SQL, REST, and GraphQL. It does not provide a ready-made dashboard interface, so teams build or connect the presentation layer separately. [Cube project]
What is Cube.js?
The Cube project describes Cube Core as “the open-source semantic layer.” In practice, it sits between data systems and the tools people use to analyze data. Rather than implement the same business metric separately in a dashboard, workbook, and application, a team can define it in Cube’s data model and expose it to multiple consumers. [Cube project]
Cube Core is headless: it provides the data model and interfaces, not a finished dashboard UI. The official project identifies BI tools, custom applications, and AI agents as consumers. A visualization tool or application still needs to render charts, tables, and other user-facing experiences. [Cube project]
How does Cube.js work?
- Connect a data source. Cube connects to SQL data systems. The project lists Snowflake, Databricks, BigQuery, Presto, Amazon Athena, and Postgres; its learning hub also covers systems such as Redshift, ClickHouse, DuckDB, Trino, MySQL, MS SQL, and Oracle. Confirm the current connector documentation for the exact source and behavior you need. [Cube project] [Cube learning hub]
- Define the data model. Specify measures, dimensions, relationships, and joins so that consumers share the same business definitions. This helps avoid embedding separate versions of metric logic in each downstream tool. [Cube project] [Cube learning hub]
- Set access rules and performance behavior. Cube’s learning materials describe row- and column-level permissions, sensitive-data masking, in-memory caching, and configurable pre-aggregations. The appropriate rules depend on the data and deployment; use documentation matching your Cube version for configuration details. [Cube learning hub]
- Expose the model to consumers. BI tools can connect through SQL, while applications and other clients can use SQL, REST, or GraphQL interfaces. Choose the interface and presentation layer that fit the consuming tool. [Cube project]
How do you use Cube.js for dashboards?
Cube Core supplies the governed data layer behind a dashboard, not the dashboard itself. A typical design connects a source to Cube, defines shared metrics and dimensions, applies permissions, and then connects a BI tool or a custom front end to query the model. This separates the definition of a metric—such as what counts as revenue—from how a particular chart displays it. [Cube project] [Cube learning hub]
#1 Best Overall
For a BI tool, SQL access can make the semantic model available through a familiar query interface. For a custom analytics application, REST or GraphQL can connect the model to the app’s own interface. Check the relevant tool and connector documentation for supported features and configuration rather than assuming every integration behaves identically. [Cube project] [Cube learning hub]
What performance should you expect?
Cube includes performance mechanisms rather than a universal speed guarantee. The project README describes a built-in relational caching engine, and Cube’s learning materials cover in-memory caching and configurable pre-aggregations. Actual results depend on the source system, data model, cache setup, workload, and deployment. The official material cited here does not establish a quantified latency claim or independent benchmark. [Cube project] [Cube learning hub]
Rank #2
- Wiley
- Language: english
- Book - storytelling with data: a data visualization guide for business professionals
How do you run Cube Core safely?
You can run Cube Core locally or self-host it with Docker. The repository’s quick-start uses development mode to simplify setup, but the project explicitly warns that development mode disables important authentication protections. Do not expose a development-mode instance to the internet or use it in production. [Cube project]
Production requires deliberate authentication and deployment configuration. Official deployment documentation notes that some production configurations require Cube Store; the appropriate topology depends on the source and deployment needs. Follow the version-matched deployment guidance rather than treating a local quick-start as a production recipe. [Cube learning hub]
Cube Core versus the commercial Cube platform
Cube Core is the open-source semantic layer. The commercial product, Cube, is an agentic analytics platform built on Cube Core. According to the project, the commercial platform adds user-facing capabilities and managed operations. The data model is compatible between the two, but the additional platform features should not be mistaken for features included in the open-source core. [Cube]
| Area | Cube Core | Commercial Cube |
|---|---|---|
| Primary role | Open-source semantic layer and API service | Agentic analytics platform built on Cube Core |
| Dashboard and workbook experience | No ready-made dashboard UI | Includes workbooks and dashboards |
| Deployment | Run locally or self-host; production configuration is your responsibility | Managed deployment is listed as a platform capability |
| Additional platform capabilities | Core semantic-layer functionality | Analytics Chat, embedded analytics surfaces, role-based access control, multi-tenancy, and integrations including Tableau, Power BI, Excel, and Google Sheets |
The choice is mainly about operating model and product scope: self-manage an API-first semantic layer and choose your own interface, or evaluate the commercial platform when managed deployment and built-in analytics experiences matter. [Cube] [Cube project]
Rank #4
Who should consider Cube Core?
- Consider it if multiple BI tools or applications need consistent metric definitions and you want a shared semantic layer.
- Consider it if you need a headless data layer with SQL, REST, or GraphQL access and are prepared to choose or build the presentation layer.
- Plan carefully if you need fine-grained access control, sensitive-data masking, caching, or pre-aggregations; implementation details should be checked against your version and environment.
- Compare the commercial platform if managed deployment, built-in workbooks and dashboards, or its listed role-based access, multi-tenancy, and integrations are priorities.
The Cube learning hub lists Cube Core v1.7—“Tesseract GA, data modeling, performance”—in a changelog entry dated July 8, 2026. Check the current release notes and versioned docs before adopting particular syntax or deployment instructions. [Cube learning hub]
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
