@budhash/gomanize

Romanize Hindi (Devanagari) into readable, diacritic-free Latin script — the gomanize engine (Go) compiled to WebAssembly, with a thin JS/TS wrapper. It is the same engine as the CLI, so output is identical; nothing is reimplemented.

Live demo: https://budhash.com/gomanize

Install

npm install @budhash/gomanize

Use

import { load } from "@budhash/gomanize";

const g = await load();                       // instantiate once
g.translit("नमस्ते दुनिया");                    // "namaste duniya"
g.translit("गाना", { longVowels: true });      // "gaanaa"
g.translit("मेरे सपनों की रानी", { rerank: true }); // "mere sapnon ki rani"

CommonJS works too: const { load } = require("@budhash/gomanize");

Options

All optional, default false:

| Group | Option | Effect | |-------|--------|--------| | rule tweak | longVowels | write "aa" for every ā (गाना → gaanaa) | | rule tweak | simpleNasals | करें → karen, not karein | | rule tweak | keepMedialSchwa | disable CCV deletion (जनता → janata) | | learned | schwaModel | decision-tree schwa classifier | | learned | lexicon | prefer attested spellings | | learned | rerank | re-rank candidates with a char n-gram model |

Notes

• load() is async (instantiates the wasm) and cached; translit is synchronous after.
• The .wasm is ~1.4 MB gzipped — it embeds the schwa model, lexicon, and n-gram data.
• Runs on Node ≥ 18 and modern browsers/bundlers (the wasm + wasm_exec.js are resolved relative to the module).
• Romanization, not strict transliteration — see the project README. MIT licensed.