Use case

[lsmith77]

I am currently looking to migrate from Fasttext and since I already use spaCy, it would be enticing to use spaCy also for the language detection pipeline. But I am a bit confused about the use case. If I load a specific language (f.e. English) and then detect the language (f.e. German), then I essentially used the wrong model on the text. So it seems more obvious to me to detect the language first. Or is the idea here to handle small text portions inside a text that is known to be in one language?