Lekhani
Copyright 2026 Udgam Labs LLP

The model weights (model.q4.gz, model.json.gz) and the runner (translit.js) are licensed under the Apache License,
Version 2.0 (see LICENSE). The word list (words.bin.gz) is dedicated to the public domain under CC0 1.0
(https://creativecommons.org/publicdomain/zero/1.0/). The model was trained from random initialisation by Udgam Labs;
no third-party model weights are included.

This model was trained on word pairs from Aksharantar (AI4Bharat), whose hand-collected part is licensed CC-BY and
mined parts CC0: https://huggingface.co/datasets/ai4bharat/Aksharantar
(Madhani et al., "Aksharantar: Towards Building Open Transliteration Tools for the Next Billion Users", 2022,
arXiv:2205.03018).

IndicXlit (AI4Bharat, MIT licence) was used as a teacher to label training words:
https://github.com/AI4Bharat/IndicXlit

The word list is derived from word counts of IndicCorp v2 Hindi (AI4Bharat, CC0):
https://huggingface.co/datasets/ai4bharat/IndicCorpV2

English words given to the teacher came from the CMU Pronouncing Dictionary
(Copyright (C) 1993-2015 Carnegie Mellon University; BSD-style licence):
https://github.com/cmusphinx/cmudict