kenya_id_ocr 0.1.2

A Flutter package for scanning and extracting data from Kenyan National ID cards using Google ML Kit OCR.

kenya_id_ocr

A Flutter package for scanning and extracting data from Kenyan National ID cards (front side) using Google ML Kit — fully on-device, no API key required.

Features

• 📷 Built-in camera scanner widget with ID card overlay guide
• 🗂️ Parse ID data from an image file or bytes
• 🔍 Extracts: ID Number, Full Name, Date of Birth, District of Birth, Gender
• ⚡ On-device ML Kit OCR — works offline
• 🛠️ Handles common OCR misreads (I→1, O→0)

Platform setup

Android

In android/app/build.gradle, ensure minSdkVersion is at least 21.

Add to android/app/src/main/AndroidManifest.xml:

<uses-permission android:name="android.permission.CAMERA"/>

<!-- ML Kit model download -->
<application ...>
  <meta-data
    android:name="com.google.mlkit.vision.DEPENDENCIES"
    android:value="ocr"/>
</application>

iOS

In ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>Camera is used to scan your Kenyan National ID</string>

Usage

Camera scanner widget

import 'package:kenya_id_ocr/kenya_id_ocr.dart';

final result = await Navigator.push<KenyaIdResult>(
  context,
  MaterialPageRoute(
    builder: (_) => KenyaIdScanner(
      onResult: (r) => Navigator.pop(context, r),
      minConfidence: 0.6, // optional, default 0.6
    ),
  ),
);

if (result != null && result.isValid) {
  print(result.idNumber);   // "12345678"
  print(result.fullName);   // "KAMAU JOHN MWANGI"
  print(result.dateOfBirth);// "01.01.1990"
  print(result.district);   // "NAIROBI"
  print(result.gender);     // "M"
}

Read from an existing image file

final reader = KenyaIdReader();
final result = await reader.readFromFile(File('/path/to/id_photo.jpg'));
await reader.dispose();
print(result);

Result model

Field	Type	Description
idNumber	String?	7–9 digit ID number
fullName	String?	Full name in uppercase
dateOfBirth	String?	As printed, e.g. 01.01.1990
dateOfBirthParsed	DateTime?	Parsed DateTime
districtOfBirth	String?	District/county name
gender	String?	"M" or "F"
confidence	double	0.0–1.0 score based on fields found
isValid	bool	True if ID number + name found
rawText	String	Full OCR output for custom parsing

Notes

• Lighting and focus significantly affect accuracy. Encourage users to scan in good light.
• The rawText field is exposed so you can build custom parsing logic on top if needed.
• This package reads the front side only.

License

MIT