ironclad_ml

Secure, RAM-only Machine Learning inference for the Ironclad SDK.

Features

This plugin prevents on-device model theft by decrypting .tflite models directly into memory (RAM) and passing them to the TensorFlow Lite engine without ever writing the unencrypted model to disk.

Usage

import 'package:ironclad_ml/ironclad_ml.dart';

// Provide your decryption key dynamically (NEVER hardcode this!)
final decryptionKey = List<int>.generate(32, (i) => i);

final interpreter = await IroncladML.loadSecureModel('assets/model.tflite.enc', decryptionKey);