中文文档 | English
Supercharge your OpenClaw with industry-leading PDF and image reading capabilities powered by PaddleOCR
- Introduction
- What is PaddleOCR Skills?
- Skills Comparison
- Why It Matters
- Prerequisites
- Installation Guide
- Resources
- License
When working with complex documents containing tables, mathematical formulas, or special layouts, AI assistants often struggle to accurately extract content. PaddleOCR Skills solves this problem by integrating the power of China's leading open-source OCR engine into OpenClaw.
With just one prompt and 3 simple steps, your OpenClaw gains professional-grade document and image parsing capabilities.
PaddleOCR Skills is a collection of AI skills available on ClawHub that enables OpenClaw to process documents and images with exceptional accuracy. It offers two core skills:
Advanced document understanding that goes beyond simple text extraction. It returns complete document structure, perfectly preserving:
- Text content with formatting
- Tables with cell structure
- Mathematical formulas (with LaTeX)
- Charts and diagrams
- Complex layouts (multi-column, headers/footers)
- Reading order and document structure
Best for: Academic papers, financial reports, invoices, legal documents, multi-column layouts
ClawHub: paddleocr-doc-parsing
Fast and accurate text extraction from images and PDFs, returning structured JSON data perfect for developers.
Best for: Screenshots, photos, scans, simple text extraction, quick OCR tasks
ClawHub: paddleocr-text-recognition
| Feature | Document Parsing | Text Recognition |
|---|---|---|
| Primary Use | Complex document understanding | Fast text extraction |
| Tables | ✅ Full structure preserved | |
| Formulas | ✅ LaTeX output | ❌ |
| Charts/Diagrams | ✅ Analyzed | ❌ |
| Layout Analysis | ✅ Complete structure | ❌ |
| Speed | Moderate | Fast |
| Output Format | Markdown + JSON | JSON |
| Best For | Academic papers, reports, invoices | Screenshots, simple images |
Use Document Parsing for:
- Documents with tables (invoices, financial reports, spreadsheets)
- Documents with mathematical formulas (academic papers, scientific documents)
- Documents with charts and diagrams
- Multi-column layouts (newspapers, magazines, brochures)
- Any document requiring structured understanding
Use Text Recognition for:
- Simple text-only extraction
- Quick OCR tasks where speed is critical
- Screenshots or simple images with clear text
- When you need structured JSON output for processing
When you send PDFs or images with complex formatting to AI assistants, they often:
- Lose table structure
- Misinterpret formulas
- Scramble multi-column layouts
- Miss important formatting details
PaddleOCR Skills provides:
- Accuracy: Industry-leading OCR from PaddleOCR (80K+ GitHub stars)
- Completeness: Preserves all document structure
- Simplicity: One-prompt installation in OpenClaw
- Free Tier: Official PaddleOCR API offers tens of thousands of free pages daily
Before installing PaddleOCR Skills, you need:
- ClawHub Account - Register at clawhub.ai
- PaddleOCR API Access - Register at paddleocr.com
- Visit www.clawhub.ai and complete registration
- Navigate to Settings → Create Tokens
- Generate and copy your token (starts with
clh_)
- Visit https://www.paddleocr.com and register
- Click the API button
- Find these interfaces:
- PaddleOCR-VL-1.5 (Document Parsing)
- PP-OCRv5 (Text Recognition)
- Copy the API_URL and TOKEN for each
Note: Official free tier supports tens of thousands of pages per day!
Send this prompt to your OpenClaw (replace the bracketed values):
Please install PaddleOCR skills for me:
ClawHub Token: [Your ClawHub Token, starts with clh_]
PaddleOCR API Configuration:
- Document Parsing API: [Your Document Parsing API URL]
- Text Recognition API: [Your Text Recognition API URL]
- Access Token: [Your PaddleOCR Access Token]
Execution Requirements: Please complete the following steps automatically:
Login to ClawHub → Install dependencies → Install skills → Write configuration → Run tests → Report installation status.
That's it! OpenClaw will automatically:
- Log in to ClawHub
- Install dependencies
- Install the skills
- Configure API credentials
- Run tests
- Report installation status
- 📖 PaddleOCR Official Site
- 🔗 ClawHub Platform
- 📄 Document Parsing Skill
- 🔍 Text Recognition Skill
- 📝 Original Tutorial (Chinese)
Scan the QR code to follow our WeChat official account for more AI tips and tutorials:
This project is licensed under the MIT License - see the LICENSE file for details.
Made with ❤️ for the OpenClaw community