Hi Product Hunt 👋 I built lipimala after running into a subtle problem with transliteration: Converting Kṛṣṇa → कृष्ण is straightforward. But converting कृष्ण back cannot tell you whether the original input was: Kṛṣṇa kṛṣṇa Kr̥ṣṇa or another accepted equivalent representation. The information simply isn't present in the visible Devanagari string. lipimala separates those two concepts: • canonical transliteration • exact original-source recovery For exact workflows, it can preserve source provenance using either: 1. a structured TransliterationResult / JSON envelope, or 2. invisible checksummed Unicode-Tag metadata attached to the rendered string. It currently focuses on: • IAST / extended Indic • Devanagari • Gujarati • Plain English / Hunterian • Vedic accent preservation • direct Devanagari ↔ Gujarati conversion There are native implementations for: • Dart • JavaScript / Node.js • Python • PHP All are maintained against the same shared behavioral corpus. The goal isn't to support the largest number of scripts. Projects like Aksharamukha already do that extremely well. The goal is deterministic, Unicode-aware transliteration where preserving the exact source matters. GitHub: https://github.com/jayeshmepani/... Feedback, Unicode edge cases and weird transliteration samples are very welcome.

Lipimala : The Indic Script Converter

Exact-source Indic transliteration across 4 runtimes

Exact-source Indic transliteration across 4 runtimes