Indic Languages Python Packages

vindex

Evals for Indian-language LLM apps: script adherence, mistranslation checks, calibrated similarity and an Indic LLM judge

polywhisper

🎙️ PolyWhisper v9: efficient multilingual Indic ASR — frozen Whisper-Small + per-language LoRA for Hindi, Tamil, Telugu, Bengali, Marathi. Bengali −34.5%, Marathi −43.2% vs baseline. ONNX CPU-ready. Apache-2.0.

indicate

transliterate hindi, punjabi to english

ocr-tamil

OCR Tamil is a powerful tool that can detect and recognize text in Tamil images with high accuracy on Natural Scenes

indicorderbench

Catches multilingual voice agents placing the wrong order. τ-bench style benchmark + CI gate for voice ordering agents in English and Hinglish: checks the committed order, not the transcript.

inltk

Natural Language Toolkit for Indic Languages aims to provide out of the box support for various NLP tasks that an application developer might need

indic-asr-onnx

Helper package for using quantized versions of the Indic ASR Model by AI4Bharat.

hindiwsd

A pipeline for transliteration, spell correction, POS tagging and word sense disambiguation of Hinglish code mixed data to Hindi Devanagari script.

akshara-tokenizer

Akshara-aware tokenizer for six Brahmic scripts, with a trained SentencePiece model

anuvaad

State of the art open-source translation for Indic languages.

quanfire-multilingual-embedding

Indic-first embeddings you can actually ship. Documented sources. Weights: CC BY-SA 4.0.

savitr

MLX-accelerated Surya OCR on Apple Silicon (~3.6x faster), for electoral-roll parsing

itranslit

transliteration for indic language

loit

Lot Of Indic Tweets

hindi-normalize

Convert Hindi number words to digits and normalize Devanagari ASR/OCR text — zero-dependency

indic-error-taxonomy

Detailed error taxonomy for Indic OCR and VLM transcriptions. Go beyond Character Error Rate to pinpoint specific mistakes in conjuncts, vowel signs, nuktas, and reading order across 9 Indic scripts.

webcorpus

Generate large textual corpora for almost any language by crawling the web