This project implements a Chinese TTS engine for Android using Kotlin and JetPack Compose with TensorFlow Lite. It runs pre-trained neural network models (FastSpeech and Tacotron originally from TensorFlowTTS) on mobile devices for real-time speech synthesis. The models are converted to TFLite format for efficient mobile inference. Users download pre-converted TFLite model files and pruned TensorFlow Lite AAR libraries to build and run the app.