github.devotd.
NewAI

libretranslate/libretranslate v1.9.6 — Self-hosted machine translation API

libretranslate/libretranslate is a free and open source machine translation project that serves the Argos Translate engine as a self-hosted REST API and web UI; the latest release is v1.9.6.

Repositorylibretranslate/libretranslate ↗

libretranslate/libretranslate is a free and open source machine translation project that serves the Argos Translate engine as a self-hosted REST API and web UI; the latest release is v1.9.6.

GitHub description: Free and Open Source Machine Translation API. Self-hosted, offline capable and easy to setup.

stars
★ 17.1K
Forks
1.8K
Language
Python
License
AGPL-3.0
Latest release
v1.9.6
Read original source ↗View on GitHub ↗

libretranslate/libretranslate is a project that provides a free and open source machine translation API and web UI for translating sentences directly. The latest public release is v1.9.6, and no release notes for this version appear in the given material, so what changed cannot be determined. It is written in Python and distributed under the AGPL-3.0 license.

The core of this repository is that translation is handled on its own server instead of being handed off to an external service. The README explains that it does not depend on proprietary software such as Google or Azure, and that the translation engine is handled by the open source library Argos Translate. The repository description states that it can be self-hosted, works in offline environments, and is not difficult to install. It offers the Argos translation engine as a REST service usable for translating sentences and documents, and the same capability is provided as a web UI as well. The star count is 17131.

That said, the scope that can be verified stops there. The material contains no figures comparing translation quality against commercial engines and no list of supported languages, and no freshness, technology, or quality scores were given. The repository was created on December 19, 2020, and the momentum score assigned by the research record reaches only 3.7 out of 10. No independently reproduced performance measurements or real adoption cases were confirmed either. The AGPL-3.0 license conditions need separate review depending on how it is actually used. Translation quality and applicable scope are hard to judge from this material alone.