pdfrx_coregraphics #

CoreGraphics-backed renderer for pdfrx on iOS and macOS.

⚠️ EXPERIMENTAL: This package is in very early experimental stage. APIs and functionality may change significantly.

This plugin provides a PdfrxEntryFunctions implementation that uses PDFKit/CoreGraphics instead of the bundled PDFium runtime. It is intended for teams that prefer the system PDF stack on Apple platforms while keeping the pdfrx widget API.

Installation #

Add the package to your Flutter app:

dependencies:
  pdfrx: ^2.2.17
  pdfrx_coregraphics: ^0.1.10

Set the CoreGraphics entry functions before initializing pdfrx:

import 'package:flutter/material.dart';
import 'package:pdfrx/pdfrx.dart';
import 'package:pdfrx_coregraphics/pdfrx_coregraphics.dart';

void main() {
  PdfrxEntryFunctions.instance = PdfrxCoreGraphicsEntryFunctions();
  pdfrxFlutterInitialize();
  runApp(const MyApp());
}

After installation, use pdfrx as usual. All PdfDocument and widget APIs continue to work, but rendering is routed through CoreGraphics.

Removing PDFium Dependencies (Reducing App Size) #

By default, pdfrx bundles PDFium shared libraries for iOS and macOS even when using pdfrx_coregraphics. If you're only using the CoreGraphics backend, you can remove these PDFium dependencies to reduce your app size.

Run this command from your project root:

flutter clean # if the environment is not clean
flutter pub get
dart run pdfrx:remove_darwin_pdfium_modules

This will comment out the iOS and macOS ffiPlugin configurations in pdfrx's pubspec.yaml, preventing PDFium binaries from being bundled with your app.

To revert the changes (restore PDFium dependencies):

flutter clean # if the environment is not clean
flutter pub get
dart run pdfrx:remove_darwin_pdfium_modules --revert

After executing it, you can run flutter build or flutter run for iOS/macOS.

Limitations #

• Incremental/custom stream loading is converted to in-memory loading
• Custom font registration is not yet supported
• In document links are always xyz and zoom is not reliable (or omitted) in certain situations
• Text extraction does not fully cover certain scenarios like vertical texts or R-to-L texts so far