Skip to main content

Security & data handling

Zingle AI is designed to support enterprise security requirements. This section describes how the platform handles your code, credentials, and data, and the guarantees the product makes about production safety.

Compliance

Zingle is designed to support enterprise security requirements. Certification evidence, audit reports, and customer-specific security documentation are available under NDA where applicable. Please contact us for the current status of any specific framework.

Principles

  • Runs in your environment. In the Customer VPC model, all components run inside your own cloud account, and code, credentials, and data stay within your boundary.
  • Human in the loop. Every change to your dbt project is proposed as a pull request you review and merge. Zingle does not merge to your default branch or deploy to production on its own. See Production safety model.
  • Least privilege. Zingle accesses your warehouse and Git using credentials and roles you provision and scope.
  • Encrypted secrets. Warehouse credentials and Git tokens are encrypted at rest. See Credential handling.

In this section

PageWhat it covers
Data handlingWhat data Zingle reads, processes, and stores.
Credential handlingHow warehouse and Git credentials are stored and used.
Secrets & access tokensPlatform secrets, encryption keys, and tokens.
LLM & data privacyWhat is sent to the language model and how.
Network access requirementsWhat Zingle connects to, and required egress.
Audit logsWhat activity is recorded.
Data retentionHow long data is kept and when it is removed.
Production safety modelThe guarantees around changes to your project and warehouse.