📇 advanced_business_card_reader #

Business/visiting card reader for Flutter using Google ML Kit Text Recognition.

✅ OCR (image → text)
✅ Best‑effort parsing: Name / Company / Phone / Email / Website
✅ Works offline (on‑device)
✅ Includes example app (Camera + Gallery + File picker)

Status: Currently tested mostly with India business cards, but the package is designed to work in other countries too (UAE / USA / Saudi / etc.) via defaultIsoCountry.

🌍 Country support (Phones) #

Phone extraction uses phone_numbers_parser and supports country‑aware parsing:

• If the card contains an international number like +971… / +1… / +91…, it parses correctly.
• If the card contains local format numbers (no +countryCode), set defaultIsoCountry.

Examples #

India

final r = await BusinessCardReader.scanAutoFromFile(
  path,
  preferredScripts: const [OcrScript.latin, OcrScript.devanagari],
  defaultIsoCountry: 'IN',
);

UAE

final r = await BusinessCardReader.scanFromFile(
  path,
  script: OcrScript.latin,
  defaultIsoCountry: 'AE',
);

USA

final r = await BusinessCardReader.scanFromFile(
  path,
  script: OcrScript.latin,
  defaultIsoCountry: 'US',
);

Saudi Arabia

final r = await BusinessCardReader.scanFromFile(
  path,
  script: OcrScript.latin,
  defaultIsoCountry: 'SA',
);

Tip: If your users are in multiple regions, set defaultIsoCountry from a user profile / app setting.

🔤 OCR language/script support #

ML Kit Text Recognition v2 is script‑based and supports:

• Latin (English, most cards worldwide)
• Devanagari (Hindi/Marathi/Nepali)
• Chinese / Japanese / Korean

Important limitation (Arabic / Tamil scripts) #

Google ML Kit Text Recognition v2 (used by google_mlkit_text_recognition) does not provide an Arabic or Tamil script recognizer in this API.

• UAE / Saudi cards written in English → ✅ works (Latin)
• UAE / Saudi cards written mainly in Arabic script → ⚠️ not supported by this ML Kit recognizer (you would need another OCR engine)

📦 Install #

dependencies:
  advanced_business_card_reader: ^1.0.0

⚙️ Platform setup (optional scripts) #

Android #

If you need non‑Latin scripts, add only what you need in your app:

android/app/build.gradle (Groovy)

dependencies {
  implementation "com.google.mlkit:text-recognition-devanagari:16.0.1"
  implementation "com.google.mlkit:text-recognition-chinese:16.0.1"
  implementation "com.google.mlkit:text-recognition-japanese:16.0.1"
  implementation "com.google.mlkit:text-recognition-korean:16.0.1"
}

android/app/build.gradle.kts (Kotlin DSL)

dependencies {
  implementation("com.google.mlkit:text-recognition-devanagari:16.0.1")
  implementation("com.google.mlkit:text-recognition-chinese:16.0.1")
  implementation("com.google.mlkit:text-recognition-japanese:16.0.1")
  implementation("com.google.mlkit:text-recognition-korean:16.0.1")
}

Latin works by default; add only the extra scripts you truly need (APK size increases).

iOS #

In your app ios/Podfile (add only required pods):

pod 'GoogleMLKit/TextRecognitionDevanagari', '~> 9.0.0'
pod 'GoogleMLKit/TextRecognitionChinese', '~> 9.0.0'
pod 'GoogleMLKit/TextRecognitionJapanese', '~> 9.0.0'
pod 'GoogleMLKit/TextRecognitionKorean', '~> 9.0.0'

Then:

cd ios
pod install

✅ Usage #

Selected script #

import 'package:advanced_business_card_reader/advanced_business_card_reader.dart';

final result = await BusinessCardReader.scanFromFile(
  filePath,
  script: OcrScript.latin,
  defaultIsoCountry: 'IN',
);

print(result.data.name);
print(result.data.company);
print(result.data.phones);

Auto script (recommended) #

For India:

final result = await BusinessCardReader.scanAutoFromFile(
  filePath,
  preferredScripts: const [OcrScript.latin, OcrScript.devanagari],
  defaultIsoCountry: 'IN',
);

For most other countries (English cards):

final result = await BusinessCardReader.scanAutoFromFile(
  filePath,
  preferredScripts: const [OcrScript.latin],
  defaultIsoCountry: 'AE', // or US / SA / etc.
);

🧪 Example app #

cd example
flutter run

Example includes:

• Camera scan
• Gallery pick
• File pick (image)
• Script dropdown (Auto + scripts)
• Parsed fields + raw text view

🧠 Tips for accuracy #

• Good lighting, no blur
• Keep the card flat and fill the frame
• Cropping before OCR improves results a lot
• Always show a confirmation/edit screen (cards vary)

🗺️ Roadmap ideas (future) #

• Country auto‑detect (based on +country code / address keywords)
• Region presets (IN / AE / US / SA)
• Better name/company heuristics for different card styles

☕ Sponsor a cup of tea #

If this package saves you time, you can sponsor:

https://github.com/sponsors/nousath

📄 License #

MIT