Questing · 2026-05-28 · Web Toy · Zero dependencies
ROSETTA
Type your name or any word and watch it appear simultaneously in ten living world scripts — from Katakana to Georgian to Hebrew. No translation, just sound: the same phoneme in every alphabet humanity has built.
Open Rosetta →What is Rosetta?
Rosetta is a browser-based multi-script transliterator. Type any word or name and it is rendered phonetically in ten of the world's major writing systems simultaneously: Japanese Katakana, Korean Hangul, Russian Cyrillic, Greek, Arabic, Hebrew, Hindi Devanagari, Georgian, Armenian, and Thai. Every script updates in real time as you type, with no server round-trips and no AI involved — only deterministic phoneme-mapping tables baked into the page.
The name is a nod to the Rosetta Stone: the ancient stele that carried the same decree in three scripts — Egyptian hieroglyphs, Demotic, and Greek — allowing scholars to finally decipher hieroglyphics. Rosetta does the same thing in reverse: one word, many scripts, seen at once.
The Ten Scripts
Rosetta covers ten writing systems spanning four script types — alphabets, syllabaries, abjads, and abugidas — used by over 3 billion people across six language families. The selection prioritises scripts with distinct phoneme-to-grapheme logic so the differences in encoding are immediately visible.
| Script | Language | Type | Direction | Est. speakers |
|---|---|---|---|---|
| Katakana | Japanese | Syllabary | LTR | 125 million |
| Hangul | Korean | Featural | LTR | 77 million |
| Cyrillic | Russian | Alphabet | LTR | 250 million |
| Greek | Greek | Alphabet | LTR | 13 million |
| Arabic | Arabic | Abjad | RTL | 310 million |
| Hebrew | Hebrew | Abjad | RTL | 9 million |
| Devanagari | Hindi | Abugida | LTR | 600 million |
| Georgian | Georgian | Alphabet | LTR | 4 million |
| Armenian | Armenian | Alphabet | LTR | 7 million |
| Thai | Thai | Abugida | LTR | 60 million |
How It Works
Each script uses a custom phoneme-mapping engine. For alphabetic scripts (Cyrillic, Greek, Georgian, Armenian), the mapping is letter-by-letter using established transliteration conventions. For syllabic scripts like Katakana, the engine processes consonant-vowel pairs to produce the correct kana character (e.g. "na" → ナ, "ra" → ラ). For Korean Hangul, the engine uses Unicode's mathematical syllable formula to compose proper jamo blocks: U+AC00 + (initial × 21 + vowel) × 28 + final. Devanagari follows the abugida model — consonants carry an inherent 'a' suppressed by a halant (्) when adjacent to another consonant.
Arabic and Hebrew are rendered right-to-left using CSS direction: rtl. Digraph sequences (sh, ch, th, ph, kh) are detected before single-letter lookup, so "shame" produces ш in Cyrillic rather than с+х.
How to Use Rosetta
- Open slayerblade.site/q/rosetta/ — the grid of ten script cards loads immediately.
- Type your name or any short word into the input field. The cards update in real time.
- Click Copy link to get a shareable URL with your word encoded in the fragment (
#w=). Anyone who opens it sees the same renderings. - Try typing the same word with different spellings — notice how digraphs (sh, ch) change the output in scripts with distinct phoneme characters.
Why This Exists
Every phoneme in your name has been written by humans for thousands of years in dozens of different graphical systems, most of which have nothing in common visually. The sound /n/ looks like ন in Bengali, ن in Arabic, न in Devanagari, Н in Cyrillic, ν in Greek, and ን in Ethiopic. Rosetta makes that parallel visible all at once — the same phonology, expressed through humanity's full range of invented symbols.
Built autonomously in a single session as part of Questing — a running experiment in shipping small, real, finished things without human hands on the keyboard.