Thank you for your interest in the new Azure SDKs! We release new features, improvements, and bug fixes every month. Subscribe to our Azure SDK Blog RSS Feed to get notified when a new release is available.
You can find links to packages, code, and docs on our Azure SDK Releases page.
Release highlights
Azure AI Foundry 1.0.0 for .NET
The first stable release of the Azure AI Foundry library for .NET provides comprehensive access to Azure AI Foundry Project resources. The library enables developers to:
- Create and run AI Agents
- Manage AI model deployments
- Handle connections to Azure resources
- Upload documents and create datasets
- Manage search indexes
This release introduces important API improvements with consistent AIProject prefixing for key models (AIProjectDeployment, AIProjectConnection, AIProjectIndex), simplified client access through direct properties on AIProjectClient (replacing individual Get*Client methods), and new convenience CreateOrUpdate methods for index management. The library uses the v1 AI Foundry data plane REST APIs and supports authentication via Microsoft Entra ID for secure, keyless access.
Azure AI Search 12.2.0 for JavaScript
This release adds support for the 2025-09-01 service version with several enhancements including support for running vector queries against sub-fields of complex fields, reranker boosted scores in search results with the ability to sort on either reranker or reranker boosted scores, and improved vector compression configuration through RescoringOptions. The release also introduces LexicalNormalizer support for text analysis, a new DocumentIntelligenceLayoutSkill for skillsets, and OneLake data source connections for indexers.
Azure AI Search 11.6.0 for Python
The Azure AI Search library for Python now supports the 2025-09-01 service version, bringing parity with the JavaScript release. Key features include vector query support for complex field sub-fields, semantic ranking enhancements with reranker boosted scores, vector compression rescoring options, lexical normalizers for improved text processing, and the new DocumentIntelligenceLayoutSkill. The release also adds QueryDebugMode support for detailed vector search debugging information.
Azure Identity libraries
The following Azure Identity libraries received important performance improvements for managed identity scenarios:
When environment variable AZURE_TOKEN_CREDENTIALS is set to ManagedIdentityCredential, DefaultAzureCredential now skips the IMDS endpoint probe request and directly attempts token acquisition with full retry logic and exponential backoff.
Initial stable releases
- Client Library for .NET
- Management Library for .NET
- Management Library for Go
- Management Library for Java
- Management Library for JavaScript
- Client Library for Python
- Management Library for Python
Initial beta releases
- Management Library for .NET
- Management Library for Go
- Management Library for Java
0 comments
Be the first to start the discussion.