One of my major “sharp edges” for using BunPro vocab is how it handles extremely uncommon kanji spellings of kana-based words. In Bunpro proper, with cloze support, I find it basically mostly upside in cases where you type in the hiragana and then it will substitute it with the uncommon kanji so that you get a little bit of recognition bonus for the rare cases where you might encounter it in the wild. Cloze support is rare in vocabulary, however. This means that much of the time I’m just presented with the vocab word completely out of context, and in this case seeing it presented as the uncommon kanji spelling is disruptive to study flow and I think also bad for recognizing the word in the wild. I can’t just click the “show furigana” button for that vocab, as from what I can tell this distinction is implemented per-kanji and not per-word, so marking one uncommon kanji word to show furigana will contaminate all other words which are supposed to be spelled with that kanji. So, at present, I need to first correctly realize that the reason I’m pulling a blank is because the word is spelled unnaturally and that I need to check the furigana. This results in something different than reading real material because now it’s “kana with kanji support”. Further, if I don’t recognize that it’s a bad spelling, I’ll spend sometimes minutes wracking my brain for no benefit. Conversely, sometimes I’ll incorrectly check for the furigana of a word that is normally spelled with kanji (because I think, maybe I’m struggling with this word because it’s “one of those”), and spoil the answer for myself. Basically it just injects a ton of chaos and inefficiency in to the tool.
Ideally I think the solution would look like a toggle at the word-detail screen which can flag this word to be presented as kana only when it comes up in reviews. This would improve my study flow, and further help me recognize it in everyday material in its native form, without needing to see the kanji to make the association. I think this may also help learner understand which words ought not be used with kanji. I sometimes get feedback that I use too much kanji and thus my text messages can be difficult / unnatural to read.