Azure Document IntelligenceInstall

Summary

Analyze documents using a selected Azure Document Intelligence prebuilt model, providing documents via binary data, URL, or base64-encoded content, and extract structured content such as text, fields, tables, and key-value pairs.

Inputs

• Model (required) — Select the prebuilt Azure Document Intelligence model to use for document analysis, such as OCR, layout, invoice, receipt, ID document, contracts, and various US-specific form types.
• Input Source (required) — Specify the source of the document to analyze: binary data from a previous node, a document URL, or base64-encoded content.
• Binary Property (required) — When using binary input source, specify the binary property name containing the document data.
• Document URL (required) — When using URL input source, provide the URL of the document to analyze.
• Base64 Content (required) — When using base64 input source, provide the base64-encoded document content.
• Additional Options — Optional parameters including content MIME type (for binary or base64 input), pages to analyze (e.g., "1-3,5"), locale hint for text recognition, and output content format (text or markdown) for layout and read models.

Output shape

Three separate outputs: (1) Content output with extracted text and content summary, (2) Structured Data with extracted overall data including pages and key-value pairs, (3) Tables output containing extracted tables formatted as JSON arrays with headers and row data.

Requests are made to Azure Document Intelligence API using provided credentials; the node handles long-running operations and returns partial results per item. Errors are reported per item if Continue On Fail is enabled. Use the content output for plain text or markdown content extraction, structured data for detailed JSON data including pages and documents, and tables output for tabular data extracted from documents.