react-native-nitro-zxing

A VisionCamera v5 barcode scanner powered by zxing-cpp, implemented as pure C++ Nitro HybridObjects. Frames are decoded straight from the camera's pixel buffer (Y plane, zero-copy) - no ML Kit, no base64, no temporary files.

npm install react-native-nitro-zxing react-native-nitro-image

Requires VisionCamera Core, Nitro Modules and Nitro Image. For frame processors also install react-native-vision-camera-worklets and react-native-worklets.

import { useBarcodeScanner } from 'react-native-nitro-zxing'

const scanner = useBarcodeScanner({ barcodeFormats: ['all-formats'] })
const frameOutput = useFrameOutput({
  pixelFormat: 'yuv',
  onFrame(frame) {
    'worklet'
    const barcodes = scanner.scanCodes(frame)
    frame.dispose()
    console.log(barcodes.map((b) => `${b.format}: ${b.displayValue}`))
  },
})

Still images: await scanner.scanCodesInImageAsync(image) with a Nitro Image.

Building from source

zxing-cpp is vendored as a git submodule at cpp/zxing-core. When cloning the repository run git submodule update --init; the npm package already contains the sources.

Coordinates

Barcode.cornerPoints / boundingBox are in the upright (rotated by Frame.orientation) frame coordinate system, matching react-native-vision-camera-barcode-scanner.