react-native-vin-plate-recognizer

On-device OCR for React Native. Google ML Kit on Android, Apple Vision on iOS. No network calls.

Scan modes:

| mode | Result type | What you get | | --- | --- | --- | | 'vin' | 'vin' | 17-character VIN in value | | 'plate' | 'plate' | UAE plate in emirate, category, number | | 'default' | 'text' | Every recognized line in value, joined with newlines |

'default' does not filter. Part codes, serials, and any other text are returned as the camera read them.

Install

npm install react-native-vin-plate-recognizer react-native-nitro-modules react-native-vision-camera

Then install pods (cd ios && pod install).

Usage

import { Camera, useCameraDevice, useFrameOutput } from 'react-native-vision-camera';
import { useSharedValue } from 'react-native-reanimated';
import { VinPlateRecognizerModule, VinPlateResult } from 'react-native-vin-plate-recognizer';

const scanMode = useSharedValue<'plate' | 'vin' | 'default'>('default');

const frameOutput = useFrameOutput({
  pixelFormat: 'rgb',
  onFrame(frame) {
    'worklet';
    const result = VinPlateRecognizerModule.recognize(frame, scanMode.value);
    frame.dispose();

    if (result.type === 'text' && result.value) {
      // one recognized line per entry — a part code can be any of them
      const lines = result.value.split('\n');
    }
  },
});

pixelFormat: 'rgb' is required. Planar YUV frames are rejected with raw: "frame_not_rgb".