Azure/azure-sdk-for-java com.azure+azure-ai-contentunderstanding

1.1.0-beta.3 (2026-08-18)

Added support for service API version 2026-06-01-preview, which is the default for this beta package.
Added synchronous and asynchronous minimal and options-based analyzeInline and analyzeBinaryInline APIs that return
ContentAnalyzerInlineResponse without long-running operation polling, preserving the inline status,
AnalysisResult, and UsageDetails. See the
analysis guidance,
Sample 19, and
Sample 20.
Added AnalyzeOptions and AnalyzeBinaryOptions overloads for model deployment overrides, input truncation,
processing location, content range, and content type. See the
analysis options guidance,
Sample 01,
Sample 19, and
Sample 20.
Added analyzer workflow selection via ContentAnalyzerConfig.setWorkflow and ContentAnalyzerWorkflow. See
Sample 17.
Added semantic chunking via ContentAnalyzerConfig.setChunkingStrategy and SemanticChunkingStrategy, with
DocumentChunk results available from DocumentContent.getChunks. See
Sample 18.
Added in-page segmentation via ContentAnalyzerConfig.setAllowInPageSegments. See
Classify in-page segments.
Added signature detection via DocumentSignature and DocumentContent.getSignatures. See
Detect signatures.
Added embedded document metadata via AnalysisContent.getMetadata. See
Extract document metadata.
Added analysis diagnostics via AnalysisResult.getInfos. See
Analysis diagnostics.
Added expanded UsageDetails values for long-running and inline analysis. See the
usage guidance,
Sample 03,
Sample 19, and
Sample 20.

Preserved source and binary compatibility with the official 1.0.0 Java API while adding preview capabilities,
including the canonical Analysis* and Content*Field models and typed ContentRange binary overload.
Updated LlmInputHelper.toLlmInput front matter with detected MIME type, separate caller and service metadata,
stable warning details, nested null/empty values, and multiline YAML scalar handling. See
ToLlmInput.