paper_document_scanner

A fully custom, themeable Flutter document scanner: live camera, automatic edge detection, draggable corner crop, filters, multi-page, and optional PDF.

Detection runs on detection-only native APIs (iOS Vision, Android OpenCV) — there is no OS system-scanner UI, so every pixel is yours to restyle and localize. See the monorepo README for why the OS scanners (ML Kit / VisionKit) cannot be themed.

Install

dependencies:
  paper_document_scanner: ^0.1.0

Usage

import 'package:paper_document_scanner/paper_document_scanner.dart';

final result = await PaperScanner.open(
  context,
  options: const PaperScannerOptions(
    outputPdf: true, // also assemble a PDF
    minPages: 1,     // done button is enabled after this many kept pages
    maxPages: 0,     // 0 = unlimited
  ),
);

if (result != null && !result.isEmpty) {
  print(result.imagePaths); // processed page images, in order
  print(result.pdfPath);    // PDF (when outputPdf: true)
}

Theming & localization

Everything visual lives in PaperScannerStyle, including all labels:

PaperScanner.open(
  context,
  style: PaperScannerStyle(
    accentColor: const Color(0xFF1B998B),
    overlayStrokeWidth: 4,
    cornerHandleRadius: 16,
    labels: const PaperScannerLabels(
      cropTitle: 'Adjust the edges',
      keep: 'Keep',
      done: 'Finish',
    ),
    // Replace individual pieces when you need to:
    captureButtonBuilder: (context, onCapture, busy) => MyShutter(onCapture, busy),
    statusPillBuilder: (context, label, style) => MyStatusPill(label),
    pageThumbnailBuilder: (context, page, onTap, style) =>
        MyThumbnail(page: page, onTap: onTap),
    cornerHandleBuilder: (context) => const MyHandle(),

    // Or replace whole chrome/action regions:
    cameraTopChromeBuilder: (context, controller, style, onCancel, onDone) =>
        MyScannerTopBar(controller, onCancel, onDone),
    cameraBottomChromeBuilder: (
      context,
      controller,
      style,
      onCapture,
      onReview,
      onCycleFlash,
      onOpenFilters,
      onToggleAutoCapture,
      capturing,
      flashIcon,
    ) =>
        MyScannerBottomBar(
          controller: controller,
          onCapture: onCapture,
          onReview: onReview,
          flashIcon: flashIcon,
        ),
    cropActionsBuilder: (context, controller, style, onRetake, onKeep) =>
        MyCropActions(onRetake: onRetake, onKeep: onKeep),
  ),
);

State is driven by a plain PaperScannerController (a ChangeNotifier) — no bloc/flutter_bloc dependency, so it never version-locks your app.

Permissions

Add a camera usage description for your platforms:

iOS — ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>Scan documents with the camera.</string>

Android — android/app/src/main/AndroidManifest.xml (the plugin also declares it, but keep your own copy explicit):

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

Android minSdk 21, iOS 13.0+. OpenCV adds ~10 MB per ABI — prefer an app bundle or ABI splits.

Output

PaperScanResult { List<String> imagePaths; String? pdfPath; }. Image paths point to processed (cropped + filtered) JPEGs in the app cache.