---
id: py3langid
version: "0.3.0"
license: BSD
license_treatment: permissive
maintenance: dormant
---

# py3langid — Fork of the language identification tool langid.py, featuring a modernized codebase and faster execution times.

License: permissive · Maintenance: dormant · Downloads: 437.9K/mo

## What it is and what it does

py3langid is a modernized fork of langid.py that detects which of 97 languages a piece of text is written in. It returns both the ISO 639-1 language code and a confidence score. The package is designed to be a drop-in replacement for the original, with significant performance improvements: imports are about 30% faster, model loading is 25-30x faster, and classification itself is 5-6x faster on paragraphs.

The tool works as both a Python library and a command-line utility. It uses a pre-trained statistical model built from diverse sources (JRC-Acquis, ClueWeb 09, Wikipedia, Reuters RCV2, Debian i18n) and requires only numpy as a dependency. It supports language subsetting, probability normalization, and can be deployed as a WSGI web service.

Use it for:

- Automatically categorize user-generated content by language for multilingual applications or content management systems.
- Filter or route text documents to language-specific processing pipelines in NLP workflows.
- Identify the dominant language in mixed-language documents or social media posts for downstream analysis.
- Build language detection into a command-line data pipeline using the standalone langid tool.
- Normalize confidence scores to probabilities for machine learning feature engineering or confidence thresholding.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Identifies the language of text in one of 97 languages using a pre-trained statistical model, with optimized speed for Python 3.6+.

Yes, if you need fast language detection across 97 languages with minimal dependencies. The dormant maintenance status is not a blocker—the package is stable, has no known vulnerabilities, and the last commit is recent enough. Install if you want a lightweight, performant drop-in replacement for the original langid.py; skip if you need active development or support for languages outside the pre-trained set.

## Install

pip install py3langid
uv add py3langid
poetry add py3langid

## Installing py3langid

Before you install: Low friction: pure Python wheel with only numpy as a runtime dependency. Dormant maintenance (last release 787 days ago, last commit 2024-11-22), but stable and archived repository shows no active development pressure.

License in practice: BSD-3-Clause permissive license allows commercial and private use with minimal restrictions; attribution and license notice required in distributions.

Quickstart:

pip install py3langid

import py3langid as langid
lang, prob = langid.classify('This text is in English.')
print(lang, prob)  # ('en', -56.77429)

Requires Python >= 3.8 and numpy; model loading on first use may take a moment.

Verify before relying:

- Whether the 30% import speedup and 25-30x model loading speedup claims are reproducible in current environments.
- Accuracy consistency across the 97 supported languages and whether results degrade on mixed-language or code-heavy text.

## Package facts

- License: BSD (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: dormant
- Downloads: 437.9K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

language detection, language identification, detect text language, multilingual text classifier, langid fork python, identify language from text, 97 language detector, language-detection, nlp-preprocessing, cli-tool

[View on SkillFed](https://skillfed.io/packages/py3langid) · [View on PyPI](https://pypi.org/project/py3langid/)