High-Accuracy Indian Language Detection in Node.js: Introducing Glossa

Building applications tailored for the Indian subcontinent requires navigating a unique and complex linguistic landscape. India is home to dozens of official languages written in a variety of distinct scripts — and frequently mixed with English in everyday digital communication.

For developers building content pipelines, search indices, routing mechanisms, or moderation workflows, accurately detecting which Indian language is being used can be a surprisingly steep challenge. Many standard global language detection packages fall short, particularly when handling different languages that share the exact same script (such as Hindi and Marathi, or Bengali and Assamese).

To bridge this gap, glossa-nodejs is a high-accuracy, zero-dependency language detection library specifically built for Indian languages in modern Node.js (20+) environments.

Why Language Detection in the Indian Context is Difficult

Global language detection libraries typically rely on character set or script block identification. While this approach functions well when distinguishing Tamil from Punjabi, it completely falls apart for same-script pairs.

Consider these two sentences:

• Hindi: मैं ठीक हूँ आप कैसे हैं
• Marathi: मी ठीक आहे तू कसा आहेस

Both sentences are written utilizing the Devanagari script. If a library only looks at character blocks, it cannot accurately differentiate between them. A similar problem exists for Bengali and Assamese, which share the Bengali script.

glossa-nodejs addresses this by combining script analysis with deterministic grammar and n-gram signals to achieve 93%+ accuracy across a dense multilingual dataset of over 1,100 cases.

Key Features at a Glance

• 14 Supported Languages: Detects Assamese, Bengali, English, Gujarati, Hindi, Kannada, Maithili, Malayalam, Marathi, Odia, Punjabi, Tamil, Telugu, and Urdu.
• Same-Script Disambiguation: Deterministic logic handles tough pairs like Hindi/Marathi and Bengali/Assamese.
• Code-Mixed Input Support: Smoothly identifies primary Indic intent even when text is interleaved with Latin/English tokens (e.g., “hello नमस्ते आप कैसे हैं”).
• Zero Dependencies: Operates entirely locally with no external runtime dependencies and no mandatory external model downloads.
• Production-Ready: Packaged for CommonJS with complete TypeScript declarations bundled out of the box.

Getting Started

1. Installation

Install the package via npm:

Bash

npm install glossa-nodejs

2. Basic Language Code Detection

For standard lookups, you can import detectLanguage to instantly retrieve an ISO-style language code.

JavaScript

const { detectLanguage } = require(‘glossa-nodejs’);

console.log(detectLanguage(‘வணக்கம் நீங்கள் எப்படி இருக்கிறீர்கள்’));

// => ‘ta’ (Tamil)

console.log(detectLanguage(‘السلام علیکم آپ کیسے ہیں’));

// => ‘ur’ (Urdu)

If the input is empty, contains only punctuation, or lacks a strong linguistic signal, it gracefully returns ‘unknown’.

3. Detection with Confidence Scores

When building automated pipelines, you often want a threshold check before triggering down-stream logic. detectLanguageWithConfidence returns both the language and a confidence interval between 0 and 1.

JavaScript

const { detectLanguageWithConfidence } = require(‘glossa-nodejs’);

const result = detectLanguageWithConfidence(‘আমি বাড়ি থেকে এসেছি’);

console.log(result);

// => { language: ‘bn’, confidence: 0.99 }

4. Code-Mixed Input & Same-Script Disambiguation

The engine handles complex inputs, like colloquial code-mixed strings or fine-grained script checks.

JavaScript

const { detectLanguage } = require(‘glossa-nodejs’);

// Code-mixed (English + Hindi)

console.log(detectLanguage(‘hello नमस्ते आप कैसे हैं’));

// => ‘hi’

// Same-script separation (Devanagari script)

console.log(detectLanguage(‘मैं ठीक हूँ आप कैसे हैं’)); // => ‘hi’ (Hindi)

console.log(detectLanguage(‘मी ठीक आहे तू कसा आहेस’)); // => ‘mr’ (Marathi)

Advanced Configurations: The Reusable LanguageDetector

For high-throughput applications requiring distinct routing rules, instantiation of a custom LanguageDetector allows you to lock down parameters such as minimum confidence thresholds or restricted language scoping.

JavaScript

const { LanguageDetector, DetectorConfig, Language } = require(‘glossa-nodejs’);

// Configure a custom detector

const config = DetectorConfig.builder()

.allowedLanguages([Language.Hindi, Language.Marathi]) // Ignore all other languages

.minConfidence(0.60)

.minTextLength(3) // Measured in UTF-8 bytes

.build();

const detector = new LanguageDetector(config);

console.log(detector.getSrcLanguage(‘मी ठीक आहे’));

// => ‘mr’

Running from the Terminal (CLI Tool)

glossa-nodejs ships with a globally accessible command-line interface, making it perfect for shell scripts, system analytics, or quick diagnostic testing on text blocks.

You can run it on-the-fly using npx:

Bash

npx glossa-nodejs “नमस्ते आप कैसे हैं”

hi

For advanced CLI formatting:

Bash

Output with confidence metrics

glossa — confidence “Hello world”

en 0.9000

Output structured JSON

glossa — format json “નમસ્તે તમે કેમ છો”

{“text”:”નમસ્તે તમે કેમ છો”,”language”:”gu”}

Stream input line-by-line via stdin

printf “வணக்கம்\nالسلام علیکم\n” | glossa — lines

ta

ur

Architectural Constraints & Considerations

When implementing glossa-nodejs in a production pipeline, keep these design considerations in mind:

Data Footprint: No runtime network calls are made. All parsing logic runs completely locally on your compute node, keeping network overhead at zero.

The Maithili/Short-Text Constraint: Maithili shares the Devanagari script with Hindi and Marathi. Without an massive external semantic model, extremely short Maithili phrases can easily be misclassified as Hindi. For highly accurate user-facing flows, treat mai classifications as best-effort unless dealing with extended text lengths.

Open Source and Contributions

The tool is written by Suraj Kumar Jha and published under the permissive MIT License. If you are handling localized search indexing, backend content classification, or building chatbot engines tailored for Indian users, glossa-nodejs provides a fast, lightweight, and isolated runtime solution.

• NPM Registry: npmjs.com/package/glossa-nodejs
• GitHub Repository: github.com/surajJha/glossa-nodejs