PDF OCR Parse

---

name: pdf-ocr-parse description: "Extract text from scanned PDFs using Tesseract OCR. Supports multiple languages, page selection, DPI control, and word-level bounding boxes."

---

PDF OCR Parse

What It Does

Rasterises each selected page of a PDF at the given DPI, then runs Tesseract OCR on each page image. Returns per-page text with confidence scores, and optionally per-word bounding boxes.

When to Use

Extract text from scanned PDF documents

OCR invoices, receipts, or legacy documents in PDF format

Extract digits-only data (invoice amounts) with char_whitelist

Process multi-language documents

Required Inputs

Provide one of:

url — URL to a scanned PDF

base64_pdf — base64-encoded PDF

Multipart upload with file field

Authentication

Send your API key in the CLIENT-API-KEY header.

Get your free API key at https://pdfapihub.com. Full API documentation is available at https://pdfapihub.com/docs.

Use Cases

Scanned Invoice Processing — OCR scanned PDF invoices to extract text for accounting systems

Legacy Document Digitization — Convert old scanned paper documents into searchable text

Insurance Claims — Extract text from scanned claim forms and medical documents

Legal Discovery — OCR scanned legal documents for full-text search and review

Multi-Language Documents — Process documents in Hindi, French, German, etc. with language-specific models

Form Digitization — Extract filled field values from scanned paper forms

Tesseract Configuration

| Param | Default | Description | |-------|---------|-------------| | lang | eng | Language code(s), + separated | | psm | 3 | Page segmentation mode (0–13) | | oem | 3 | OCR engine mode (0=legacy, 1=LSTM, 3=default) | | dpi | 200 | Rasterisation DPI (72–400) | | char_whitelist | — | Restrict to specific characters |

Example Usage

curl -X POST https://pdfapihub.com/api/v1/pdf/ocr/parse \
  -H "CLIENT-API-KEY: your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://pdfapihub.com/sample-pdfinvoice-with-image.pdf",
    "pages": "1-3",
    "lang": "eng",
    "dpi": 300,
    "detail": "words"
  }'