Context Management
Manage context to improve agent performance across your workspace.
- Available on the Team and Enterprise plans.
- Users need the Admin or Manager role to access context management features.
The Context Management tab gives you tools to improve agent performance by shaping the context and guidance agents rely on when. You can add and refine context using guides, endorsements, and semantic models, then validate changes in the testing workbench before publishing them live.

Guides
The Guides tab is where you can provide unstructured context that helps agents interpret questions and respond appropriately across your workspace. Read more about Guides.
Endorsements
Endorsements indicate which data assets are trusted in your workspace. Endorsed assets are prioritized when agents answer questions and act as a clear visual signal to end users that the data has been reviewed and approved by workspace Admins or Managers.
You can manage endorsements for multiple asset types, including:
- Schemas and tables in data connections
- Projects in your workspace
- Semantic models

You can also manage asset statuses directly on this page. See Endorsed statuses for more details.
Managing endorsements via API
In addition to the Context Studio UI, you can programmatically manage endorsements using the Hex API. This is useful for bulk updates or integrating endorsement management into your existing workflows. The following endpoints are available:
UpdateSemanticProject— Update status for datasets and views in semantic modelsUpdateDataConnectionSchema— Update status for warehouse databases, schemas, and tablesUpdateProject— Update status for projects
Semantic models
Semantic models define shared metrics, dimensions, and relationships used across Hex. When configured, agents can use these definitions to answer questions more consistently.
Context Studio shows how semantic models contribute to agent behavior and gives you another entry point into authoring and managing semantic models. To learn how to create and manage semantic models, see Semantic modeling.
