PaddleOCR.js

English | 简体中文

Official browser OCR SDK and demo for PaddleOCR.

Project structure

Path	Role
packages/core/	Browser SDK sources; published to npm as @uzen/paddleocr-js
demo/	Vite demo app that depends on the SDK

Features

• Text detection + recognition (OCR) pipeline
• Document orientation classification (0°/90°/180°/270°) with auto-rotation
• Document unwarping / flattening preprocessing
• Text line orientation classification (0°/180°) for pre-recognition correction
• Web Worker mode support
• Fully local browser inference — no cloud upload required
• Custom model support (PP-OCRv5, PP-OCRv6)
• YAML pipeline configuration
• Visualization utilities for rendering OCR results
• Multi-language support (Chinese, English, Japanese, etc.)

Quick start

import { PaddleOCR } from "@uzen/paddleocr-js";

const ocr = await PaddleOCR.create({
  lang: "ch",
  ocrVersion: "PP-OCRv5"
});

const result = await ocr.predict(blob);
console.log(result[0].items);

Local development and demo

npm install
npm run dev:demo

Other common commands:

npm run build
npm run test
npm run typecheck
npm run check

Documentation

Topic	Link
Architecture	architecture.md
Development	development.md
Monorepo conventions	monorepo.md
SDK package README	packages/core/README.md

Acknowledgements

• PP-OCR for the OCR models
• ONNX Runtime Web for browser-side ONNX inference
• OpenCV.js for browser-side image processing