@napi-rs/system-ocr

OCR library that uses system provided API. VisionKit on macOS, Media OCR on Windows.

node example/index.js

Example	Result

pnpm add @napi-rs/system-ocr
yarn install @napi-rs/system-ocr
npm install @napi-rs/system-ocr

import { recognize } from '@napi-rs/system-ocr'

const result = await recognize('path/to/image.png')

for (const line of result.lines) {
  console.log(line.text, line.confidence, line.boundingBox)
}

result.lines follows the system OCR engine's reading order. Each bounding box uses normalized input-image coordinates (0 to 1) with its origin at the top-left. The formatted result.text is preserved for compatibility and is not guaranteed to equal the line texts joined together. Line confidence is 1.0 on Windows, matching the existing result confidence convention.

import { recognize, OcrAccuracy } from '@napi-rs/system-ocr'

const image = await fetch('https://example.com/image.png')

const result = await recognize(image, OcrAccuracy.Accurate, ['fr', 'zh-cn'])

Huge thanks to:

• win-ocr-rs
• mac-system-ocr