DavidVentura/offline-translator

Use Firefox Translation Models for on-device translation on Android

What it solves

It provides a way to translate text, images, and documents (PDF/ODT) on Android devices without requiring an internet connection, ensuring privacy and availability in offline environments.

How it works

The app uses on-device machine learning models for all its core functions. Users download language packs once, and the app then performs translations locally. It leverages Firefox's translation models for text, PaddleOCR for image text recognition, and various engines like MNN and slimt for inference. It also includes a built-in dictionary based on Wiktionary data and supports multiple TTS (Text-to-Speech) voices.

Who it’s for

Android users who need translation services while offline, people prioritizing data privacy by avoiding external servers, and developers who want to integrate translation capabilities into other apps via the provided API.

Highlights

  • Fully Offline: No requests are sent to external servers after language packs are downloaded.
  • Multimodal Translation: Supports plain text, images, and PDF/ODT documents.
  • Broad Language Support: Covers a wide array of global languages including Chinese, Arabic, Hindi, and Japanese.
  • Integrated Tools: Includes automatic language detection, transliteration for non-latin scripts, and a built-in word dictionary.
  • Developer API: Exposes an API for other Android applications to request translations.

Related

  • Project
  • Project
  • Project
  • Project
  • Project