ocrCN — Multi-Platform Chinese OCR Skill

中文文档

Extract text from images and documents with Chinese-optimized OCR. All backends work in China, no VPN required.

Platform Comparison

Quick Start

# Install PaddleOCR (default backend — free, no API key)
pip install paddlepaddle paddleocr

# Extract text from an image
python skills/ocrCN/scripts/ocr.py image.png

Natural language (Claude Code):

Just tell Claude what you want:

Extract the text from this screenshot
Convert this invoice image to text
Extract the info from this ID card
Recognize this table and output as markdown
Batch OCR all images in this folder

Backend Selection

PaddleOCR (Default — Free, Offline)

Baidu's open-source OCR engine. The gold standard for Chinese OCR. No registration, no API key, unlimited usage.

• 15+ specialized models: detection, recognition, table, formula, seal
• GPU acceleration (CUDA)
• PDF + image input
• JSON output with bounding boxes and confidence

Baidu AI OCR (Cloud API)

Baidu's cloud OCR with 50+ specialized models for Chinese document types.

• ID card, business license, VAT invoice, bank card
• Driver's license, vehicle license, license plate
• Table, seal/stamp, formula, handwriting

Tencent Cloud OCR (Cloud API)

Tencent's OCR with smart structural analysis — auto-detects tables, forms, paragraphs.

Alibaba Cloud OCR (Cloud API)

Alibaba's DocMind — document layout analysis, separates text, tables, and figures.

EasyOCR (Free, Offline)

General-purpose multilingual OCR. Lighter than PaddleOCR but less accurate for Chinese.

Usage Examples

# Chinese + English mixed
python skills/ocrCN/scripts/ocr.py --lang ch_en mixed.png

# ID card (Baidu AI)
BAIDU_APP_ID="xxx" BAIDU_API_KEY="xxx" BAIDU_SECRET_KEY="xxx" \
python skills/ocrCN/scripts/ocr.py --platform baidu --type id_card id.jpg

# VAT invoice (Baidu AI)
BAIDU_APP_ID="xxx" BAIDU_API_KEY="xxx" BAIDU_SECRET_KEY="xxx" \
python skills/ocrCN/scripts/ocr.py --platform baidu --type invoice fapiao.jpg

# Table to Markdown (Tencent Cloud)
TENCENT_SECRET_ID="xxx" TENCENT_SECRET_KEY="xxx" \
python skills/ocrCN/scripts/ocr.py --platform tencent --type table --format markdown table.png

# Complex layout (Alibaba DocMind)
ALIBABA_ACCESS_KEY_ID="xxx" ALIBABA_ACCESS_KEY_SECRET="xxx" \
python skills/ocrCN/scripts/ocr.py --platform ali --layout paper.png

# Handwriting recognition
python skills/ocrCN/scripts/ocr.py --handwriting note.jpg

# Batch + visualize
python skills/ocrCN/scripts/ocr.py --visualize screenshots/*.png -o results/

Config File

Create ~/.ocrCN.json for defaults:

{
  "backend": "paddle",
  "lang": "ch",
  "min_confidence": 0.8
}

Priority (highest first): CLI args > project config (.ocrCN.json) > user config (~/.ocrCN.json) > env vars > built-in defaults.

Installation

Claude Code

# Plugin marketplace (recommended)
/plugin install ocrCN@365-skills

Or tell your coding agent:

help me to install https://github.com/Agents365-ai/ocrCN.git

# Manual install
git clone https://github.com/Agents365-ai/ocrCN.git ~/.claude/skills/ocrCN/

OpenClaw

git clone https://github.com/Agents365-ai/ocrCN.git ~/.openclaw/skills/ocrCN/

SkillsMP

Discover and install at skillsmp.com.

Requirements

• Python 3.8+
• pip
• (Optional) CUDA GPU for PaddleOCR acceleration

Support

If this project helps you, consider supporting the author:

Author

Agents365-ai

• Bilibili: https://space.bilibili.com/441831884
• GitHub: https://github.com/Agents365-ai

License

CC BY-NC 4.0 — free for non-commercial use. Commercial use requires permission.