How each case style decides which letters change
Upper and lower case are simple only in English. This page uses the letter rules of the text language you pick, the same rules your browser applies for that language. In Turkish, i and ı are two different letters, so “istanbul” becomes “İSTANBUL” with its dot, and “ISPANAK” lower-cases to “ıspanak” without one. German “Straße” becomes “STRASSE”, because ß traditionally has no capital form. Greek has two lower-case sigmas, and “ΟΔΥΣΣΕΥΣ” comes back as “οδυσσευς”, with the final ς. Pick English for these words and you get the English result, which a Turkish or Greek reader would see as a spelling mistake.
Sentence case lower-cases the text and capitalizes the first letter after a full stop, question mark, exclamation mark or line break, plus the English pronoun I. A full stop needs a space after it, so 3.5 and maxscanner.app are not taken as sentence ends. Capitalized Case capitalizes the first letter of every word. Both keep words that already have a capital inside them (NASA, iPhone, McDonald) and links or file names as you typed them, because a converter cannot know whether “apple” is a fruit or a company. When the whole text is in capitals there is nothing to learn from it, so every word is re-cased.
Title Case is where style guides differ. AP and Chicago both capitalize the first and last word of a title and the first word after a colon. AP style, used by newsrooms and many websites, lower-cases articles and the conjunctions and prepositions of three letters or fewer, and capitalizes longer ones: “Notes From the Field”. Chicago style, used for books and essays, lower-cases prepositions of any length, the conjunctions and, but, for, or and nor, and the words to and as: “Notes from the Field”. They also treat yet and so differently: AP writes “Small yet Mighty”, Chicago “Small Yet Mighty”. Words that usually belong to a verb, such as up in “Set Up”, are capitalized; if one of them is a plain preposition in your title, change it by hand.
The code styles treat each line as one name. Spaces, hyphens, underscores, dots and existing capitals split the words, so “XMLHttpRequest” gives xml_http_request in snake_case and xmlHttpRequest in camelCase. Apostrophes are dropped, so “isn’t” becomes isnt. Accented letters are kept unless you tick Remove accents, which turns “Crème brûlée” into creme_brulee for systems that accept only ASCII names. Scripts without capital letters, such as Hindi, Arabic or Chinese, keep their letters and only gain the separators.