Paddleocr Text Recognition

PaddleOCR wraps the PP-OCRv5 API to pull text from images, PDFs, and screenshots. You pass it a URL or local file path, it returns line-level strings with optional bounding boxes. The docs are thorough about when to use it: grab text from scans and photos, skip it for tables or complex layouts where you need structure preserved. Strong on CJK and handwritten text according to the description. Requires API credentials from paddleocr.com, runs via uv with inline PEP 723 dependencies. Output defaults to timestamped JSON files in temp unless you explicitly pipe to stdout. If you need raw text extraction without wrestling with layout reconstruction, this does the job cleanly.

skills.sh installs: 609
GitHub stars: 92.5K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add paddlepaddle/paddleocr --skill paddleocr-text-recognition --agent claude-code

Installs into .claude/skills of the current project.

Files

PaddleOCR Text Recognition

When to Use This Skill

Use this skill for:

Extract text from images (screenshots, photos, scans)
Extract text from PDFs or document images when the goal is line/box-level text
Extract text from URLs or local files that point to images/PDFs

Do not use for:

Documents with tables, formulas, charts, or complex layouts — use Document Parsing instead

Usage

Basic OCR

From URL:

paddleocr api \
  --model_type ocr \
  --file_url "https://example.com/image.png"

From local file:

paddleocr api \
  --model_type ocr \
  --file_path "./document.pdf"

Common Options

# With specific model
paddleocr api \
  --model_type ocr \
  --model PP-OCRv5 \
  --file_path "./report.pdf"

# Disable preprocessing (faster, for flat/well-oriented images)
paddleocr api \
  --model_type ocr \
  --file_path "./document.pdf" \
  --use_doc_unwarping False \
  --use_doc_orientation_classify False

# Save result to file
paddleocr api \
  --model_type ocr \
  --file_url "https://..." \
  --output result.json

SKILL.md on GitHub