argostranslate

What it is and what it does

Argos Translate is a Python library for offline neural machine translation built on OpenNMT's CTranslate2. It lets you translate text between language pairs by downloading pre-trained model packages and running them locally without external API calls. The library can automatically chain translations through intermediate languages when a direct model isn't available—for example, translating Spanish to French by routing through English if you have es→en and en→fr models installed.

You interact with it through Python code, a command-line interface, or by building on top of it (LibreTranslate is an example web service built on Argos Translate). The package handles model discovery, installation, and management. It supports GPU acceleration via CTranslate2 when configured with the ARGOS_DEVICE_TYPE environment variable, and depends on several NLP libraries (spacy, stanza, sacremoses, sentencepiece, minisbd) for preprocessing and tokenization.

Use it for:

• Translate user-generated content in a web or desktop application without sending data to a third-party translation API
• Build a command-line tool that translates documents or text snippets locally on demand
• Integrate translation into a data processing pipeline where you need to preserve privacy or avoid API rate limits
• Experiment with neural machine translation models and language pivoting strategies for research or prototyping
• Deploy a translation service on edge devices or servers where external API access is unavailable or undesirable

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Argos Translate is an open-source offline neural machine translation library that translates text between many language pairs using OpenNMT's CTranslate2 backend, with support for automatic language pivoting when direct translations are unavailable.

Yes, if you need offline neural translation and can manage model downloads. The library is actively maintained, permissively licensed, and has low install friction. Main considerations: you must download language models before translating (not included in the package), and the seven NLP dependencies add complexity. No known security vulnerabilities. Best suited for applications requiring privacy, offline operation, or avoiding API dependencies; less ideal if you need quick setup with minimal configuration.

Install

argostranslate on PyPI

pip install argostranslate

uv add argostranslate

poetry add argostranslate

Installing argostranslate

Before you install

Low install friction with a pure Python wheel distribution. Actively maintained with recent commits and stable production status. Seven runtime dependencies (ctranslate2, spacy, stanza, sacremoses, sentencepiece, minisbd, packaging) are well-established NLP libraries, though the full translation pipeline adds moderate complexity.

License in practice

Licensed under MIT (permissive), so you can use, modify, and distribute the package freely in commercial or private projects with minimal restrictions.

Quickstart

pip install argostranslate

import argostranslate.package
import argostranslate.translate

argostranslate.package.update_package_index()
available_packages = argostranslate.package.get_available_packages()
package_to_install = next(
    filter(
        lambda x: x.from_code == "en" and x.to_code == "es",
        available_packages
    )
)
argostranslate.package.install_from_path(package_to_install.download())

translatedText = argostranslate.translate.translate("Hello World", "en", "es")
print(translatedText)

Requires downloading and installing language model packages (.argosmodel files) before translation; GPU acceleration requires setting ARGOS_DEVICE_TYPE environment variable and CTranslate2 GPU support.

Verify before relying

• Whether language model downloads are cached locally or re-downloaded on each use
• Typical latency and throughput for translation operations on CPU vs GPU
• Memory footprint when multiple language models are installed simultaneously
• Quality degradation when pivoting through intermediate languages vs direct translation

Package facts

Evidence: argostranslate-1.11.0-py3-none-any.whl

Tags

More Linguistic packages