Extract Document
AI-generatedSummary
Upload a PDF, invoice, receipt, or image document (from a specified binary field) to extract structured data in various formats (JSON, Markdown, Text, or Spreadsheet) using AI and OCR processing. Configure extraction details like output format, processing mode, page range, extraction schema (for JSON or Spreadsheet), and whether to wait for the processing to complete.
Inputs
- Binary File (required) — Whether to take the document from a binary field in the input data
- Input Binary Field (required) — Name of the binary property containing the document file (PDF or image) to process
- Output Format (required) — Desired format of the extracted data: JSON, Markdown, Text, or Spreadsheet
- Extraction Schema — JSON Schema defining fields to extract; required if output format is JSON or Spreadsheet
- Processing Mode (required) — Trade-off between processing speed and accuracy: Fast, Balanced, or Accurate
- Page Range — Specific pages to process (e.g., '1-5,10'); processes all pages if empty
- Wait for Completion (required) — Whether to wait for the document processing to complete before continuing
- Max Wait Time (Seconds) — Maximum time to wait for processing completion if waiting is enabled
Output shape
a single JSON object containing the document processing status, extracted data, and metadata if waitForCompletion is true; otherwise, returns an object with the document ID.
If Wait for Completion is true, the operation polls for processing status up to Max Wait Time seconds and returns the full extracted data. If false, returns immediately with the document ID for later retrieval. For JSON and Spreadsheet output formats, an extraction schema is required specifying the structured fields to extract. Errors or timeouts are returned as node operation errors.