The naïve transliterators are more accurate, while this goes for simplicity. If you don't know about Lojban: it's a constructed language, and it assimilates loanwords and foreign-language names by respelling them with its own phonemic writing system (this site's name would be rendered as la .outkykiz. or la .otkukiz. or along those lines). For the tools here, what is important to recognize are three non-intuitive sound values: c stands for an [ʃ]-like sound, j stands for an [ʒ]-like sound, x stands for an [x]-like "strong h" sound, and y stands for an [ə]-like vowel sound. There are no digraphs. It's a bit loose with the h letter: it's ordinarily not used in Lojban, but I have chosen to use it in some circumstances to transliterate a few letters of Arabic and Hebrew (contrasting with x), and to mark aspiration like in Devanagari.
Supported scripts: Cyrillic (Russian-style), Greek (both ancient-like and modern-like, check the options), Hebrew, Arabic, Japanese kana, various Indic scripts, Armenian.