Azure AI Content Understanding SDK for Python
Extract content from documents, images, audio, and video with Azure AI
Transform unstructured media files into structured, searchable content for RAG applications and automated workflows. Use prebuilt analyzers or create custom field extraction models with the Azure AI Content Understanding SDK.
下载技能 ZIP
在 Claude 中上传
前往 设置 → 功能 → 技能 → 上传技能
开启并开始使用
测试它
正在使用“Azure AI Content Understanding SDK for Python”。 Analyze a research paper PDF using prebuilt-documentSearch
预期结果:
Document analysis complete. Extracted 15 pages of markdown content including 3 figures, 2 tables, and 47 paragraphs. Content structured with proper headings and citations preserved.
正在使用“Azure AI Content Understanding SDK for Python”。 Transcribe a 30-minute meeting recording
预期结果:
Audio transcription complete. 1,247 words across 89 timestamped phrases. Key topics detected: project timeline, budget review, resource allocation. Full transcript available with speaker segmentation.
正在使用“Azure AI Content Understanding SDK for Python”。 Extract invoice fields from vendor PDF
预期结果:
Custom analyzer extracted: vendor_name='Acme Corporation', invoice_total=15420.50, invoice_date='2026-01-15', line_items=[{description: 'Software License', amount: 12000}, {description: 'Support', amount: 3420.50}]
安全审计
安全This skill contains documentation for the Azure AI Content Understanding SDK, an official Microsoft Azure service. Static analysis scanned 0 files with 0 security issues detected (risk score: 0/100). The SKILL.md file provides usage examples for document, image, audio, and video analysis using legitimate Azure SDK methods. No executable code, network calls, or dangerous patterns present.
质量评分
你能构建什么
RAG Document Indexing
Convert PDF documents, research papers, and technical manuals into markdown format for retrieval-augmented generation systems.
Meeting Intelligence
Transcribe recorded meetings and webinars with speaker timestamps for searchable meeting notes and action item extraction.
Invoice Processing Automation
Extract vendor names, invoice totals, and line items from supplier invoices using custom field extraction models.
试试这些提示
Analyze this PDF document using the prebuilt-documentSearch analyzer and return the extracted markdown content. Document URL: {url}Transcribe this audio file and provide all phrases with their start and end timestamps. Audio URL: {url}. Format output as: [start - end]: textAnalyze this video to extract key frames, transcript phrases, and generate a summary. Video URL: {url}. Return: 1) Key frame descriptions with timestamps, 2) Full transcript, 3) Executive summaryCreate a custom analyzer with fields: {field_schema}. Then analyze this document: {url} and extract values for each defined field. Return results as structured JSON with field names and extracted values.最佳实践
- Use begin_analyze with AnalyzeInput for all analysis operations and await poller.result() for completion
- Access extracted content via result.contents[0] and check the content.kind property for type-specific handling
- Prefer async client with azure.identity.aio for high-throughput scenarios requiring concurrent analysis jobs
避免
- Do not call analyze() directly - always use begin_analyze() which returns a poller for long-running operations
- Avoid accessing result.fields without first verifying the analyzer was configured with a field_schema
- Do not use sync client for batch processing workflows that require analyzing multiple files concurrently