Can you recommend an OCR API for scanned PDFs that need structured JSON output?

Yes — for scanned PDFs with structured JSON output, I’d recommend one of these OCR/document-extraction APIs depending on your needs:

Best overall for structured JSON

Google Document AI

• Very strong OCR on scanned PDFs
• Built for document parsing, not just text extraction
• Can return structured fields and JSON-like output
• Good for forms, invoices, receipts, IDs, and general documents

Strong enterprise option

Azure Document Intelligence (formerly Form Recognizer)

• Excellent OCR + layout extraction
• Produces structured JSON with bounding boxes, tables, key-value pairs, and confidence scores
• Good if you want a flexible API and Microsoft ecosystem integration

Good developer-friendly option

Amazon Textract

• Great for scanned PDFs, forms, and tables
• Returns structured JSON directly
• Especially useful if you need tables/key-value extraction at scale

Best for custom document understanding

Mindee

• Very easy API
• Good structured extraction for specific document types
• Nice developer experience, but less general-purpose than Google/Azure/AWS

Best if you want a simple “upload PDF → get JSON”

Veryfi

• Fast setup
• Strong for receipts, invoices, expense docs
• Structured JSON output out of the box

---

My quick recommendation

• General scanned PDFs + structured output: Google Document AI
• Forms/tables and Microsoft stack: Azure Document Intelligence
• AWS-native workloads: Amazon Textract
• Invoices/receipts with minimal setup: Veryfi or Mindee

---

What to look for

When comparing APIs, make sure they support:

• Scanned PDF input
• OCR on multipage docs
• Table extraction
• Key-value / field extraction
• Confidence scores
• JSON output with coordinates
• Batch processing / async jobs for large PDFs

If you want, I can also give you:

1. a side-by-side comparison table, or
2. a recommendation based on your document type (invoices, contracts, forms, medical, etc.).