The search is sensitive to diacritics. The following characters have to be typed in Unicode (or copied from here):
ˀ ˁ á ã ᵃ ḏ ḏ̣ é ẽ ᵉ ɛ έ ə ə́ γ ḥ ḫ ʰ í ⁱ ḳ ľ ḷ ó ö ö́ ᵒ ɔ ɔ́ š ṣ ṣ̌ ŝ ṯ ṭ ú ʸ ẓ̂ ž
The base is searchable by regular expressions. For example, ^ can be used for indicating the starting position within the string, \b can be used for indicating word boundaries, and . (a point) can be used as the wildcard.