Skip to main content

Configure AI Settings

Dremio Cloud provides managed AI functionality to help you generate labels, wikis, and SQL statements. The AI features are powered by enterprise-grade models and are managed by Dremio to ensure optimal performance and security.

Enable AI Features​

  1. Click in the side navigation bar to go to the Project Settings page.
  2. Select Preferences in the settings sidebar.
  3. Enable the Generative AI toggle.

Supported AI Providers​

Dremio Cloud provides managed AI services with support for multiple enterprise-grade model providers. The following providers are available:

OpenAI​

  • Models: GPT-4o, GPT-4 Turbo, and other enterprise-grade models
  • Connection: Managed by Dremio Cloud
  • Use Cases: Text-to-SQL, label generation, wiki generation, semantic search

AWS Bedrock​

  • Models: Various foundation models available through AWS Bedrock
  • Connection: Managed integration with your AWS environment
  • Use Cases: Enterprise AI workloads with AWS-native models

Azure OpenAI​

  • Models: OpenAI models deployed through Azure OpenAI Service
  • Connection: Managed integration with your Azure environment
  • Use Cases: Enterprise AI workloads with Azure-native deployment

Default AI Provider​

Unlike self-hosted Dremio deployments where you configure model providers directly, Dremio Cloud manages AI provider configuration for you. The default provider and available models are:

  • Fully managed: No direct configuration required
  • Enterprise-grade: Uses the latest models optimized for data tasks
  • Secure: Data is processed according to Dremio's security standards
  • Multi-provider: Support for OpenAI, AWS Bedrock, and Azure OpenAI

To modify AI provider settings or request access to specific models, contact your Dremio administrator or Dremio support.

Available AI Features​

Once enabled, Dremio Cloud's AI functionality provides:

Text-to-SQL Experience​

Convert natural language queries into SQL directly within the SQL Runner. See Text-to-SQL Experience for more information.

Generated Labels and Wikis​

Automatically generate descriptive labels and documentation for your datasets. See Generating Labels and Wikis for more information.

Use AI-powered search to discover datasets and entities across your project. See Semantic Search for more information.

Data Privacy and Security​

  • Neither Dremio nor OpenAI trains models using customer data
  • Data is secured with user credentials and processed according to enterprise security standards
  • Schema information and metadata may be used as context for AI operations
  • All AI features are audited by the same practices and policies as other Dremio Cloud features

For detailed privacy information, see Data Privacy.