@oomfware/lang-detect #

natural language detection library.

npm install @oomfware/lang-detect

usage #

call initialize() once to load model weights via fetch(), then use detect() synchronously on any text. results are sorted by probability, highest first.

import { initialize, detect } from '@oomfware/lang-detect';

await initialize();

const results = detect('the quick brown fox jumps over the lazy dog');
console.log(results[0]); // -> ['eng', 0.98]

mixed-script text returns detections for each script present, with probabilities scaled by proportion:

const results = detect('Hello Мир');
// -> [['eng', ...], ['rus', ...]]

supported languages #

50 languages across Latin, Cyrillic, Arabic, Devanagari, CJK, and unique-script families.