Deep-OCR icon

Deep-OCR

AI-powered OCR for invoices, receipts, contracts, IDs, bank statements and more — returns structured JSON ready to use

Deep-OCR

AI-generated

Summary

Upload a document from an n8n binary property to Deep-OCR, request extraction using the selected document type, and return the completed extraction as JSON.

Inputs

  • Binary Property (required) — Name of the input binary property containing the document; defaults to data. Accepts PDF, PNG, JPG/JPEG, WebP, or TIFF files up to 100 MB.
  • Document Type — Extraction schema to request; defaults to Generic. Choose a specific type such as Invoice, Receipt, Contract, Bank Statement, Delivery Note, ID Document, Payslip, or Purchase Order, or use Handwriting or Full-Text transcription. Auto-Detect asks the API to classify the document first and costs more; Full-Text is billed at 10× a structured extraction.

Output shape

One output item per successfully processed input item, with extracted fields in its JSON object.

The output spreads the API's result.content fields at the top level and adds document_type and metadata. The node submits an extraction job and polls until it completes or fails. With Continue On Fail enabled, a failed item instead returns { "error": "..." }.

Examples

Example 1: Invoice PDF

Upload a PDF on binary property data and choose Invoice to extract vendor, line items, taxes, totals, and payment terms.

Example 2: Unknown document type

Choose Auto-Detect when the document type is genuinely unknown; the API classifies it before extraction, at higher cost.

Example 3: Full transcription

Choose Full-Text to receive the document transcribed as Markdown; this option costs 10× a structured extraction.

Links

Discussion