Two schemes, two jobs
Romanizing Arabic means picking the right trade-off. If a human needs to read the result — a citation, a name in a bibliography, a place on a map — you want ALA-LC, the Library of Congress standard with its familiar diacritic letters. If a machine needs to process it — search, sort, store in an ASCII field — you want Buckwalter, a strict reversible mapping. This tool gives you both from the same input.
How it works
Each scheme is a character lookup table applied across your text.
- ALA-LC maps consonants to readable Latin equivalents (
ث → th,ص → ṣ,ع → ʿ) and renders short vowels when harakat are present. A shadda doubles the preceding consonant, reproducing the gemination. - Buckwalter assigns every Arabic letter and diacritic a single distinct ASCII symbol (
ح → H,ش → $,ء → ', shadda →~). Because the mapping never merges two glyphs, it round-trips perfectly back to Arabic.
Tips and notes
For names and bibliographic work, ALA-LC is almost always what a publisher or library catalogue expects. For building a search index, deduplicating a corpus, or feeding text into a pipeline that chokes on non-ASCII, Buckwalter is the safer choice precisely because it is lossless. If your source is fully vocalised you will see richer ALA-LC output, since the short vowels carry through; strip the harakat first if you only want the consonantal romanization.