AI Models & Platforms
Starburst Introduces Enterprise Context Layer for AI Agents

Key PointsContext Studio centralizes business rules and metrics so agents stop inferring meaning…
View more
- Context Studio centralizes business rules and metrics so agents stop inferring meaning from scattered sources.
- The private preview marks all context as certified while steward-controlled review remains under development.
- Context Studio entered private preview on October 5, 2026 before the Enterprise Context Layer umbrella announcement.
Starburst Data introduced its Enterprise Context Layer at Launch Point 2026 on November 18, 2026, presenting Context Studio and AI-ready Data Products as a way to give AI agents governed business meaning alongside access to distributed enterprise data.
Starburst says the latest advances to AIDA, its AI Data Assistant, expand the questions agents can answer, the data they can access and the workflows they support, while customers’ own agents can use the same governed context. The intended result is more accurate, consistent and traceable answers from distributed enterprise data.
How Context Studio Works
Enterprise schemas locate and describe data without reliably encoding the rules behind a business question. A case record marked as pending may or may not count as open, and a column name cannot explain how net exposure is calculated or who qualifies as an active customer. Starburst’s Context Studio documentation says this leaves agents to infer meaning that may be scattered across data products, code repositories, metadata stores, unstructured sources and analysts’ knowledge, producing different answers to the same question.
Context Studio consolidates glossary terms, metrics, business rules, synonyms, join keys and data products into one governed store. Its Context Studio as Code API receives context from repositories and pipelines, while an out-of-the-box Galaxy harvester derives entities from existing Starburst Galaxy data products and query history. The store tracks each entity’s origin. Because every source push describes that connection’s complete current state, an entity omitted from a later push is removed from the store.
A named steward can review and certify entries through the Context Studio interface or certification API, and agents prioritize certified context. During private preview, however, all context enters as certified while Starburst develops that capability further. AIDA, Claude and other agents supporting Model Context Protocol (MCP), a standard for connecting AI applications to Galaxy data, can use the Context Studio MCP server. They apply certified definitions, metrics, rules and join keys when writing and running queries, with Starburst Galaxy applying the individual user’s access policy each time.
Preview History and Access Requirements
The Launch Point announcement is broader than a first release of Context Studio. Starburst Galaxy’s release notes date private-preview support to October 5, 2026, listing the Galaxy harvester, Context Studio as Code API and MCP access to certified context. The newer development is the Enterprise Context Layer umbrella and the introduction of AI-ready Data Products within it, not the start of Context Studio availability.
Context Studio remains in private preview. Organizations need a Galaxy account enrolled in the preview, administrator access to that account and access to the agent or AI client they plan to connect. Starburst requires interested customers to contact their account team for access and current preview status. Usage limits are disclosed during onboarding and may change.
For enterprises running several agents, the design centralizes business vocabulary and query logic while preserving user-level data controls. The key execution issue is the certification workflow: shared context can already be distributed across MCP clients, but the preview marks every item as certified, leaving steward-controlled review as a central piece still under development.












