Offline_SOS_System

Pub.dev Package: Link

GitHub Repository: Link

Imagine getting into a serious car crash in a remote area—a mountain pass, a highway dead zone, or a rural road with zero cell signal.

You open your safety app, or its automated background trigger fires... only to hang indefinitely because it relies on a cloud API to process sensor data or verify the crash.

That single point of failure bugged me for months. Emergency safety features shouldn’t depend on a stable 5G connection. If an engine can detect a crash instantly via onboard physics, our software should be able to do the same on-device.

So, I built and open-sourced ** offline_sos_system**—a pure Dart, 100% offline crash detection engine powered by on-device TensorFlow Lite.

Most existing Flutter solutions for safety or impact detection suffer from one of three issues:

G-force > X

spikes, leading to massive false-positive rates (like dropping your phone on a table or hitting a pothole).I wanted a solution that was pure Dart/Flutter at the developer layer, handled complex multi-axis motion patterns via Edge AI, and never made a single network request.

The package handles the entire pipeline locally on the device:

tflite_flutter

.Here is how simple it is to initialize and listen for crash events in Flutter:

dart
import 'package:offline_sos_system/offline_sos_system.dart';

void main() async {
  WidgetsFlutterBinding.ensureInitialized();

  // Initialize the offline SOS engine
  final sosEngine = OfflineSosSystem();
  await sosEngine.initialize();

  // Listen to real-time crash detection events
  sosEngine.crashStream.listen((CrashEvent event) {
    if (event.isCrashDetected) {
      print('CRASH DETECTED!');
      print('Confidence Score: ${event.confidence}');
      print('Impact Force: ${event.gForce}G');

      // Trigger your app's local emergency protocols here
    }
  });

  // Start monitoring sensor telemetry
  await sosEngine.startMonitoring();
}