@thomasfuchs It’s a welcome improvement, but I‘m sad that it doesn’t use the correct „quotation marks“ after recognizing the language.