How do I use an OCR API for multi-page PDFs and TIFFs?

To use an OCR API for multi-page PDFs and TIFFs, the basic flow is usually:

1. Upload the file to the OCR service
2. Tell the API to process every page
3. Poll or wait for completion if it’s asynchronous
4. Fetch the extracted text and/or structured results

Typical API workflow

1) Send the document

Most OCR APIs accept:

• application/pdf
• image/tiff
• sometimes ZIPs or image arrays

For multi-page documents, the API should treat each page as a separate image internally.

2) Specify page handling

Look for parameters like:

• pages=all
• page_range=1-10
• async=true
• multi_page=true

If you don’t specify anything, many APIs default to all pages.

3) Choose output format

Common OCR outputs:

• plain text
• JSON with page-by-page text
• hOCR / ALTO XML
• searchable PDF
• bounding boxes / coordinates

For multi-page documents, JSON is often easiest because it includes page numbers and per-page confidence.

4) Handle async processing

Large PDFs/TIFFs are often processed asynchronously:

• POST /ocr returns a job ID
• GET /ocr/{job_id} returns status
• when done, download results

5) Parse page results

Expect structure like:

{
  "pages": [
    { "page": 1, "text": "..." },
    { "page": 2, "text": "..." }
  ]
}

---

Example cURL request

curl -X POST "https://api.example.com/ocr" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "file=@document.pdf" \
  -F "language=eng" \
  -F "pages=all" \
  -F "output=json"

---

Example async flow

Submit job

curl -X POST "https://api.example.com/ocr/jobs" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "file=@scan.tiff" \
  -F "output=json"

Response:

{ "job_id": "abc123", "status": "queued" }

Poll status

curl -X GET "https://api.example.com/ocr/jobs/abc123" \
  -H "Authorization: Bearer YOUR_API_KEY"

When complete:

{
  "status": "completed",
  "result_url": "https://api.example.com/results/abc123.json"
}

---

Important tips for PDFs/TIFFs

• Check page limits: some APIs cap pages or file size
• Prefer TIFF for scans if you need high-quality page images
• Use searchable PDF output if you want text embedded in the document
• Split very large files if the API has timeouts
• Preserve page order in your application
• Convert rotated pages if OCR accuracy is poor

---

If you’re building against a specific OCR API

I can give you:

• a Python example
• a JavaScript example
• a cURL example
• or the exact steps for a specific service like Google Vision, AWS Textract, Azure Document Intelligence, or Tesseract-based APIs

If you want, tell me which OCR API you’re using and I’ll tailor the request format.