Skip to content
BSWEN
Blog

How to OCR Scanned PDFs with OpenDataLoader: 80+ Languages Including CJK

Cowrie
Cowrie
Dev @ Bswen

Problem

Many PDFs are scanned images with no selectable text. Standard PDF parsers output nothing useful. OCR solutions exist, but the accurate ones are either commercial (Azure, Google Cloud) or require GPU (marker, Surya). If you work with multilingual documents — Korean reports, Japanese papers, Chinese contracts — the problem gets harder.

Solution: OpenDataLoader Hybrid Mode with OCR

OpenDataLoader’s hybrid mode includes a full OCR pipeline. It runs locally on CPU without GPU, supports 80+ languages, and works with poor-quality scans at 300 DPI+.

Install hybrid mode
pip install "opendataloader-pdf[hybrid]"

Enable OCR

Start the backend with --force-ocr. This makes the triage system route scanned pages to the OCR pipeline.

Terminal 1: Start backend with OCR
opendataloader-pdf-hybrid --port 5002 --force-ocr
Terminal 2: Process scanned PDFs
opendataloader-pdf --hybrid docling-fast scanned_doc.pdf -o output/

Multi-Language Support

Combine language codes with commas:

Korean + English OCR
opendataloader-pdf-hybrid --port 5002 --force-ocr --ocr-lang "ko,en"

Supported language codes include: en, ko, ja, ch_sim (simplified Chinese), ch_tra (traditional Chinese), de (German), fr (French), ar (Arabic).

Python API

Python with OCR
import opendataloader_pdf


opendataloader_pdf.convert(
    input_path=["scanned_doc.pdf"],
    output_dir="output/",
    format="markdown,json",
    hybrid="docling-fast"
)

Key Points

  • All processing is local — no cloud API calls, no data leaving your environment
  • No GPU required — runs on CPU
  • 80+ languages including full CJK support
  • Works with poor-quality scans at 300 DPI+

Summary

In this post, I showed how to OCR scanned PDFs with OpenDataLoader hybrid mode. The key point is that you get free, local OCR for 80+ languages including CJK — no cloud dependency, no GPU, no data exfiltration risk.

Final Words + More Resources

My intention with this article was to help others share my knowledge and experience. If you want to contact me, you can contact by email: Email me

Here are also the most important links from this article along with some further resources that will help you in this scope:

Oh, and if you found these resources useful, don’t forget to support me by starring the repo on GitHub!

Comments

Tags: