NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #3782 most downloaded on PyPI
Microsoft Corporation Azure AI Content Understanding Client Library for Python
Last release 1 months ago
11 Aug 2026
Ships fairly regularly
a new release about every 6 weeks
Nearly every release is documented
notes for 3 of 3 stable releases
Nothing withdrawn
no release was ever pulled
8 months old
7 releases · first in 2026
One column per month.
Added support for selecting 2025-11-01 or 2026-06-01-preview through the api_version keyword argument on ContentUnderstandingClient; this beta package
Added support for selecting 2025-11-01 or 2026-06-01-preview through the api_version keyword argument on ContentUnderstandingClient; this beta package defaults to 2026-06-01-preview. See the README service API version examples.
Added support for 2026-06-01-preview service. Features included:
analyze_inline and analyze_binary_inline. See sample_analyze_inline.py and sample_analyze_binary_inline.py.ContentAnalyzerConfig.workflow to ContentAnalyzerWorkflow.AGENTIC. See sample_create_analyzer_workflow.py.ContentAnalyzerConfig.allow_in_page_segments. See sample_classify_in_page_segments.py.ContentAnalyzerConfig.chunking_strategy with SemanticChunkingStrategy and reading DocumentContent.chunks. See sample_analyze_chunking.py.DocumentSignature and DocumentContent.signatures. See sample_detect_signatures.py and sample_analyze_configs.py.AnalysisContent.metadata and to_llm_input. See sample_extract_document_metadata.py and sample_to_llm_input.py.AnalysisResult.infos. See sample_analysis_diagnostics.py.UsageDetails, available from AnalyzeLROPoller.usage, AnalyzeAsyncLROPoller.usage, and ContentAnalyzerInlineResponse.usage. See sample_analyze_invoice.py, sample_analyze_inline.py, and sample_analyze_binary_inline.py.to_llm_input caller dictionary from metadata to custom_metadata; it is emitted under a nested customMetadata: front-matter block.gpt-5.2 and text-embedding-3-large, including prebuilt analyzer deployment aliases and a model retirement schedule note..github/skills/ for user feedback on iterative custom-analyzer authoring in VS Code:
cu-sdk-author-analyzer — author and refine a custom document analyzer for a single document type (layout extraction → schema drafting → validation → batch test → agent review → refine cycle).cu-sdk-author-analyzer-classify-route — author and refine a classify-and-route pipeline for mixed-document packets (e.g. invoice + bank statement + loan application in one PDF), with per-category review of both the outer classifier descriptions and each inner schema's field descriptions.Filtered service-emitted LLMStats: telemetry entries from the rendered rai_warnings front matter.
LLMStats: telemetry entries from the rendered rai_warnings front matter.to_llm_input page markers from <!-- page N --> to <!-- InputPageNumber: N --> and avoided duplicate marker injection when the service markdown already includes InputPageNumber markers.Added to_llm_input helper function that converts AnalysisResult objects into LLM-friendly text with YAML front matter and markdown content. Supports d
to_llm_input helper function that converts AnalysisResult objects into LLM-friendly text with YAML front matter and markdown content. Supports documents, audio/video, and classification hierarchies.sample_create_analyzer_with_labels (sync + async): added an analyze step (calls begin_analyze on the newly created analyzer to extract MerchantName / TotalPrice from a sample invoice when training data is configured), a DEMO MODE banner when no training data is configured, a field-schema verification banner, and try / finally cleanup so the analyzer is deleted even if creation fails.Code | Docs
Communication Phone Numbers
azure-communication-phonenumbers
Added usage property on AnalyzeLROPoller and AnalyzeAsyncLROPoller to surface billing and token consumption details (UsageDetails) returned by the RES
usage property on AnalyzeLROPoller and AnalyzeAsyncLROPoller to surface billing and token consumption details (UsageDetails) returned by the REST API.Removed _models.pyi stub file that caused type checkers (pyright, mypy) to only resolve 10 of 51 model classes, hiding types like AnalysisResult and A
_models.pyi stub file that caused type checkers (pyright, mypy) to only resolve 10 of 51 model classes, hiding types like AnalysisResult and AnalyzerDefinition. The .value property type information is now provided via TYPE_CHECKING class redeclarations in models/_patch.py.GA release of Azure AI Content Understanding client library for Python
ContentField subclass (e.g., StringField, NumberField) now exposes a value property with a type appropriate to that subclass (e.g., str for StringField, float for NumberField)The following API changes were made from the preview SDK (1.0.0b1) to the GA SDK to align with Azure SDK for Python design guidelines:
AnalyzeInput → AnalysisInput, AnalyzeResult → AnalysisResult, MediaContent → AnalysisContent, MediaContentKind → AnalysisContentKindAnalysisInput.input_range → content_rangebegin_analyze inputs parameter is now a required keyword argument (previously optional); parameter order changed to inputs, model_deployments, processing_locationbegin_analyze_binary input_range keyword renamed to content_range; parameter order changed to content_range, content_type, processing_locationsample_create_analyzer_with_labels for label-based training (sync + async)Initial release of Azure AI Content Understanding client library for Python
ContentUnderstandingClient for analyzing documents, audio, and video contentYour coding agent can read these notes before it upgrades. Set up the MCP server →