ocr

Recognize Chinese and English text in images and PDFs using PaddleOCR.

Updated Mar 10, 2026

Search Tags:#ocr #pdf #python #image #multilingual #text-extraction #paddleocr

Ready to run

npx skills add https://github.com/xongtao/openclaw-backup --skill ocr-xongtao

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ocr
Source: https://github.com/xongtao/openclaw-backup/tree/main/skills/ocr-python
Command: npx skills add https://github.com/xongtao/openclaw-backup --skill ocr-xongtao

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires paddleocr, PyMuPDF, and includes scripts (resource) components.

What problem does it solve?

OCR recognises text from images and PDFs, turning non-searchable content into usable text.

Core Features & Use Cases

Chinese and English text recognition via PaddleOCR.
Image and PDF text extraction, with page-by-page handling for multi-page documents.
Simple integration in Python workflows for downstream processing or indexing.

Quick Start

Run OCR on a file path to extract and output the recognized text for immediate use.

Frequently Asked Questions about ocr

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I extract text from a scanned PDF using Python?▼

To extract text from scanned PDFs, this Skill processes each page with PaddleOCR, recognizing Chinese and English text and aggregating the results into editable output. It uses PyMuPDF to handle multi-page documents seamlessly within a Python workflow.

Can I use PaddleOCR to convert images with Chinese text to editable text?▼

Yes, PaddleOCR supports multilingual text extraction, specifically recognizing Chinese and English text within images. It converts non-searchable image content into usable text through simple Python integration for immediate downstream processing.

Does this OCR approach handle multi-page PDFs page by page?▼

Yes, this OCR approach handles multi-page PDFs by processing each page individually and aggregating the recognized text. It applies PaddleOCR to scanned documents and images, returning the complete extracted text for the entire document.

What's the best way to turn non-searchable images into searchable text?▼

The best way to turn non-searchable images into searchable text is using PaddleOCR via a Python script. It recognizes text in images and PDFs, converting non-searchable content into usable text quickly and outputting it for indexing or further processing.

Do I need PyMuPDF to extract text from PDFs with PaddleOCR?▼

Yes, PyMuPDF is required as a dependency to process PDF files alongside PaddleOCR. PyMuPDF handles the PDF page extraction, allowing PaddleOCR to apply Chinese and English text recognition to each page individually.

Are there limitations when using PaddleOCR for multilingual text extraction?▼

PaddleOCR handles multilingual text extraction focusing on Chinese and English recognition. It is optimized for scanned documents and images, but complex layouts or unsupported languages may affect the accuracy of the output formatting.