Google Unveils Offline Translation Device Gemma Translator
The Gemma Translator demonstrates a fully local, open-source translation device that works without internet access, built by a small team at Google Creative Lab.
Reporting from 1 source: GIGAZINE.
Google has announced the Gemma Translator, a translation device that operates offline. It uses the Gemma 4 E2B AI model on a Raspberry Pi 5 to translate speech locally, with transcription and speech synthesis handled by Moonshine's AI. The device's design files and source code are open-sourced.
The Gemma Translator is a device developed by a small team at Google Creative Lab. It is equipped with a microphone and speaker; when you speak into the microphone, it is translated into the target language and audio is played from the speaker. It can also display the text before and after translation on a monitor.
It is equipped with the single-board computer Raspberry Pi 5, and runs Google's lightweight AI model Gemma 4 E2B on Google's edge device inference framework LiteRT-LM to perform translation processing. The transcription before translation and speech synthesis after translation are executed locally using Moonshine's AI. Since all processing is done locally, you can converse with multilingual speakers even in places without internet access.
Google's AI agent, Antigravity, was used in the development of the Gemma Translator. The design files and source code for the Gemma Translator are published on GitHub.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.