At Cube Rollup, our first-ever in person user conference, we announced several new features for Cube Cloud during the keynote. These features include our next-generation data modeling engine that we’ve given the codename Tesseract, Cube Visual Modeler, Data Access Policies, and Cube Copilot. Additionally, we announced new capabilities for Semantic Catalog, and its promotion to being generally available to Cube Cloud users. With these exciting new features, users will see performance improvements, improved developer experience, and easier collaboration.
Gain Performance Enhancements with Next-Generation Data Modeling
Cube began as an open-source project built entirely in JavaScript, but over the years, it has evolved, with nearly 60% of its codebase now in Rust, driven by the SQL API and Cube Store. Rust has proven to be an ideal language for data platforms, allowing Cube to integrate key libraries like DataFusion, arrow-rs, and egg. Now, Cube is introducing its Next-Generation Data Modeling Engine, code-named Tesseract, which will further increase Rust’s presence and introduce powerful new capabilities like multi-stage calculations, while enhancing data model performance.
Tesseract aims to optimize Cube’s data modeling by improving SQL code generation, compiling models faster, reducing memory footprint, and supporting complex SQL queries, such as period-over-period comparisons and percentage of total calculations. These improvements enable users to easily handle more advanced data modeling tasks, reducing the need for manual post-processing or complex SQL workarounds.
A key feature of Tesseract is multi-stage calculations, which simplify nested aggregations, such as year-to-date (YTD) calculations, and make advanced metrics like percentage of total or year-over-year growth easier to define directly within the data model. With Tesseract now in preview and multi-stage calculations available, data engineers can upgrade to the latest version to try out these powerful new features.
Apply Granular Role-Based Controls with Data Access Policies
While Cube has had strong access control abilities for many years, we have heard from users that it can be difficult to understand exactly how to use the provided features such as visibility settings, query rewriting, and dynamic data modeling to achieve the exact results that they desire. If you’ve spent any time working as a database administrator, you’ve no doubt spent a lot of time working through issuing the correct privileges for an object to a user, or a specific role and wondering if you’ve done it correctly. Add to this complexity our integrations with third party identity providers and things become even more challenging.
With Data Access Policies, now in preview, we aim to make it easier for users to define and visualize policies that govern access to data with an easy-to-use user interface in Cube Cloud. Data Access Policies will enable users to maintain access control within the cubes and views definitions that make up the data model.
By introducing Data Access Policies we believe it will be easier to develop a governance model and maintain it, supporting developer productivity and reducing the risk of security misconfiguration leading to unintended data exposure. Use the new access_policy parameter within a cube or view definition to provide a declarative definition of your access policy for the object. These policies provide an expressive, yet easy to understand syntax for defining governance of data assets.
Increase Engineering Productivity with AI in Cube Copilot
Data engineers are constantly seeking tools that not only simplify their workflows but also enhance their ability to deliver robust data models to meet the demands of the business. Today, we're excited to announce the release of Cube Copilot, now in preview, a groundbreaking AI assistant designed to empower data engineers to build and manage the universal semantic layer more efficiently than ever before.
Cube Copilot assists users by providing real-time suggestions based on the current context. It can suggest code snippets, and even respond to natural language prompts in the form of comments in your code. Add a comment describing what you’d like to create in your data model code and Cube Copilot will generate the required code. Go faster and get more done with help from Cube Copilot.
Turn Clicks into Code with Cube Visual Modeler
Cube has historically been a heavily code-oriented platform. Most of us at Cube, when explaining the data modeling capabilities of Cube, use the phrase “code-first” somewhere in our explanation. Data models, defined in Javascript and YAML, and easily managed with Git integration is a hallmark of Cube. Cube comes from deep developer roots.
We recognize that organizations increasingly have an overwhelming need to be able to involve more individuals in the data modeling process. It’s a common situation that the person with the most subject matter experience on a given data domain is not comfortable with code. This exclusion is a loss for our customers, and the broader data industry.
To support broader collaboration, and open the aperture of who can contribute to the data modeling process, we are happy to announce Cube Visual Modeler, now in preview, a canvas-style shared workspace where you can visualize, create and modify data models. Replacing Data Graph, Canvas will be the place to see and edit data models directly. Users can build new cubes and views from scratch directly within Canvas in a true no-code experience.
Each change made on the Canvas interface within Cube Visual Modeler is translated into code behind the scenes. This generated code still goes through all approval workflows, maintaining consistency no matter how the data model code is developed. Even if you’re an experienced Cube developer, the new Canvas experience can help you visualize how your cubes and views are interrelated. Code-first, or no-code - Cube Cloud now supports how you and your team would like to work.
Discover Connected Data Assets with Semantic Catalog
Semantic Catalog, our unified view of connected data assets, is now generally available. This feature provides a self-documenting and continuously updated catalog based on the data models defined within Cube Cloud. Key capabilities of Semantic Catalog include unified search of connected data assets, insights into data lineage and relationships and exploration of dependent content downstream of Cube.
In addition to now being generally available, this feature has been expanded to now include column and member level lineage, allowing even greater insights into your data assets. Semantic Catalog is available in Cube Cloud Premium and above.
Conclusion
We’ve been hard at work preparing for Cube Rollup, our first user conference and we couldn’t be more excited to tell the world about all of the new features and improvements that we’re bringing to Cube. Countless hours went into nailing down a set of releases that we think are going to inspire users to tackle their biggest challenges and expand their ideas of what a universal semantic layer can be.
We’re sincerely grateful for all of the users, partners and friends who joined us to celebrate live in San Francisco, but if you couldn’t make it to the event, we’re happy to bring all the announcements to you here on the Cube blog. If you’d like more information about any of these releases, please check out the individual technical blog linked about each one. If you’d like to take them for a spin, contact sales, or review the documentation for directions to enable them.