Paste your text and every curly quote, apostrophe and prime becomes the straight version your code was expecting. One character, three hours of your afternoon — we have all been there.
Typesetters have always used curved quotation marks, because they are easier to read and they tell you which end of a quotation you are at. Typewriters could not manage it, so we got the straight ones on your keyboard instead, and those are what programming languages and file formats were built around.
Then word processors started swapping them back in as you typed, and language models learned to write from books that were set properly in the first place. So the quotes in your text are the typographically correct ones. They are simply not the ones the box you are pasting into can read.
A curly quote in source code is a syntax error, and an unusually confusing one: the line looks perfect, and the editor points at a character that appears to be exactly what it should be. Hours have been lost to that.
JSON rejects them outright. CSV files stop quoting fields the way a parser expects. Search boxes miss the phrase you are certain is in the document. Older systems and some databases turn them into question marks, and then the mistake is baked in permanently.
None of this is a problem in prose. It is a problem the moment prose has to become data.
Quotation marks come in pairs, so they are easy to spot. The apostrophe does not, and the curly one hides in every contraction you write: don’t, it’s, we’ll. One paragraph of ordinary prose can carry a dozen without a single quotation in sight.
It is the same character as the closing single quote, which is why a find-and-replace for quotation marks usually catches it by accident and a search for apostrophes usually does not.
It is also the one that breaks SQL. Whether the apostrophe is curly or straight, text going into a query belongs in a bound parameter rather than glued into the string.
No, on purpose. Going that way means guessing from context whether each mark opens or closes, and the guess is wrong often enough to matter: the apostrophe in rock 'n' roll and in the '90s curls the opposite way to what any simple rule produces. A tool that silently gets that wrong is worse than no tool, so this one only goes the direction it can get right every time.
Because in French, German, Danish, Russian and several other languages they are the correct quotation marks, not a mistake. Converting them would be like straightening out an accent. Tick the box if you are deliberately flattening text for a system that only understands ASCII.
A prime is the mark for feet and minutes, and a double prime for inches and seconds. They lean where an apostrophe curls, and they are different characters again. Almost nobody types the real ones deliberately, but they arrive with text copied out of documents that were set properly.
Yes. Everything you paste is treated the same, and for code that is nearly always what you want, since a curly quote in source is a syntax error. If you are only after the prose, convert the prose.
Quotes are usually not the only thing that came along. Em dashes and invisible characters travel in the same paste and cause the same sort of baffling error, where the line looks perfect and the parser disagrees. Run it through the Em Dash Remover and the Invisible Character Remover as well.