azure-ai-projects-dotnet

You can manage Azure AI Foundry agents in .NET by creating an AIProjectClient instance with DefaultAzureCredential, then creating agents or agent versions, running persistent threads, and polling for run completion to retrieve responses.

To run automated evaluations in Azure AI Projects, use the .NET SDK to run evaluations against uploaded datasets and integrate Azure OpenAI chat clients for model responses, which is ideal for CI workflows and model validation.

The best way to upload datasets and create indexes in Azure AI Projects using .NET is through the AIProjectClient, which supports dataset uploads, creating or updating indexes, and listing deployments within your project endpoint.

Yes, you need DefaultAzureCredential to authenticate with the Azure AI Projects SDK for .NET. You also need a project endpoint URI to initialize the AIProjectClient for managing agents, datasets, and deployments.

Yes, Azure AI Projects in .NET supports production agent orchestration and CI workflows by enabling versioned agent creation, persistent thread execution, automated evaluations against datasets, and deployment configuration.

Azure AI Projects agent threads require async polling in .NET because the SDK uses asynchronous operations to monitor run completion, ensuring your application retrieves the final agent response without blocking execution.