LanguageGuard¶

LanguageGuard restricts content to specified languages using Unicode script detection and character frequency analysis. No external model required.

Overview¶

Property	Value
Latency	<5ms
Memory	50 KB
Async	No
ML Required	No
License	Community

Supported Scripts¶

Script	Languages
Latin	English, French, Spanish, German, Portuguese
Cyrillic	Russian, Ukrainian, Bulgarian
CJK	Chinese, Japanese (Kanji), Korean (Hanja)
Arabic	Arabic, Farsi, Urdu
Devanagari	Hindi, Sanskrit, Marathi
Hangul	Korean
Hiragana/Katakana	Japanese
Greek	Greek
Thai	Thai
Hebrew	Hebrew

Usage¶

Rust¶

use oxideshield_guard::guards::LanguageGuard;
use oxideshield_guard::{Guard, GuardAction};

let guard = LanguageGuard::new("lang")
    .allow_script(Script::Latin)
    .with_action(GuardAction::Block);

let result = guard.check("Hello world");
assert!(result.passed);

Python¶

from oxideshield import language_guard

guard = language_guard(allowed_scripts=["latin"], action="block")
result = guard.check("Hello world")
assert result.passed

Configuration¶

guards:
  - type: language
    action: block
    allowed_scripts:
      - latin
      - cyrillic

Research References¶

LLM Guard Language Scanner — Language detection
Unicode Script Property — Script identification standard