How this phonetic Urdu typing tool works
Type Roman Urdu (Urdu sounded out with English letters) and the tool converts it to Urdu script as you go. It first checks each word against a built-in dictionary of very common Urdu words (hai, hain, mein, kya, shukriya and similar), so everyday words come out correctly. Anything not in that dictionary falls back to letter-by-letter phonetic conversion, using common digraphs like "kh", "gh", "ch", "sh" and "aa" for sounds that need two English letters.
The retroflex-letter convention
As a convention borrowed from common Roman-Urdu transliteration schemes, capital T, D, R and N produce the retroflex letters (ٹ, ڈ, ڑ, ں) instead of their regular dental versions - useful for words like "baRa" or "paani" written with the capital marking the harder sound. This mirrors how many Urdu speakers already type informally in Roman script across messaging apps, rather than inventing a new convention.
Why a dictionary-first approach instead of pure letter conversion
Many of the most frequent Urdu words don't transliterate cleanly letter-by-letter - checking a built-in dictionary of common words first, before falling back to phonetic rules, means everyday words like "hai", "mein", and "shukriya" come out correctly rather than as an approximate letter-salad that a native reader would need to mentally correct.
Where phonetic approximation genuinely has limits
Urdu has sounds and spelling conventions - certain vowel lengths, specific letter choices among several that sound similar - that a general phonetic ruleset can't always resolve the way a native speaker would spell a given word. Letter-by-letter conversion is a strong approximation for unfamiliar or uncommon words, not a guarantee of the exact traditional spelling.