Support TokenCredential-based Authentication for Evaluators🔐✨ #39620
Labels
Client
This issue points to a problem in the data-plane of the library.
customer-reported
Issues that are reported by GitHub users external to the Azure organization.
Evaluation
Issues related to the client library for Azure AI Evaluation
feature-request
This issue requires a new behavior in the product in order be resolved.
needs-team-attention
Workflow: This issue needs attention from Azure service team or SDK team
Service Attention
Workflow: This issue is responsible by Azure service team.
Is your feature request related to a problem? Please describe.
Currently, model configuration - see, configs.py supports only API key authentication. This limitation extends to evaluators such as the groundedness evaluator in azure-ai-evaluation SDK, forcing users to rely solely on API keys. This is suboptimal, as one of the design goals for Azure AI Foundry SDKs (and SFI) is to support secure, token-based authentication methods (Entra) over API keys. 😕
Describe the solution you'd like
It would be ideal if the SDK could support authentication via
TokenCredential
(DefaultAzureCredential/Entra-ID). This enhancement would update the model configuration to accept credential objects in addition to, or instead of, API keys. Doing so would align the azure-ai-evaluation SDK with broader Azure SDK authentication and SFI practices, offering a more secure and consistent experience for users. 🚀Describe alternatives you've considered
Additional context
Adopting token-based authentication would:
Feel free to reach out if further details or discussion is needed!
The text was updated successfully, but these errors were encountered: