Skip to content
LibreTranslate

LibreTranslate

A self-hosted machine translation tool supporting multi-language translation, usable offline with controllable data privacy

Features

Open SourceTranslation

System Requirements

16GB RAM recommended. 10GB+ storage recommended.
macOS 15+: Supports both Intel and M-series chips.
Windows 10/11: Intel/AMD GPUs supported, NVIDIA GPU recommended.
Note: For NVIDIA GPUs, install a newer driver.

Introduction

LibreTranslate is a free and open-source machine translation API that anyone can deploy on their own server. It does not rely on commercial translation services like Google or Microsoft; instead, it uses open-source technology to provide translation capabilities. It is ideal for users who want to protect data privacy, need offline translation, or avoid restrictions from third-party services.

Note: The language models downloaded during the default installation are en,zh,zt,ko,fr,ja. For a full list of supported languages, please refer to: https://docs.libretranslate.com/guides/supported_languages/

To download additional language models, use the following command, es and it represent Spanish and Italian respectively:

python download_models.py es,it

Key Features

  1. Free and Unlimited:No paid subscriptions, with no limits (number of translations) or character count.
  2. Self-hosted:Can be installed on your own computer or server, ensuring data never leaves your infrastructure.
  3. Offline Capable:Works without an internet connection once language models are downloaded.
  4. Easy to Use:Offers a web interface and API, making it simple for developers to integrate into applications.
  5. Multi-language Support:Supports translation between multiple languages (depending on installed models).

Development Team & Technology
LibreTranslate is developed and maintained by Piero Toffanin and community contributors, and is not owned by a specific commercial company. Its translation engine is powered by Argos Translate, an open-source lightweight machine translation framework that uses pre-trained open models.