Google adds on-device ASR and TTS in AI Edge Gallery

Google has introduced on-device Automatic Speech Recognition and Text-to-Speech capabilities in Google AI Edge Gallery, allowing developers to build offline voice interfaces for small language models like Gemini Nano and Gemma 4. The release includes modular tools such as ML Kit and MediaPipe Tasks, which handle audio pre-processing and hardware acceleration across Android and cross-platform devices. Additionally, the open-source infrastructure utilizes LiteRT to package models into unified .litertlm files for local deployment. Open article →